DUAL-GENE TRANSFER AND VECTOR TARGETING FOR

HEMATOPOIETIC STEM CELL GENE THERAPY

by

JUSTIN CHARLES ROTH

Submitted in partial fulfillment of the requirements

for the degree of Doctor of Philosophy

Dissertation Advisor: Dr. Stanton L. Gerson

Program in Molecular Virology

Division of Hematology/Oncology

CASE WESTERN RESERVE UNIVERSITY

January, 2006 CASE WESTERN RESERVE UNIVERSITY

SCHOOL OF GRADUATE STUDIES

We hereby approve the thesis/dissertation of

______

candidate for the ______degree *.

(signed)______(chair of the committee)

______

______

______

______

______

(date) ______

*We also certify that written approval has been obtained for any proprietary material contained therein. DEDICATION

This work is dedicated to my family, without their constant encouragement and loving support, it would not have been possible.

vi TABLE OF CONTENTS

List of Tables ...... iii

List of Figures ...... iv

Preface ...... vii

Acknowledgements ...... ix

List of Abbreviations ...... x

Abstract ...... xiv

Chapter 1: Hematopoietic Stem Cell Gene Therapy ...... 1

Stable Gene Transfer Vectors ...... 1

HSC Expansion: Inducers of Proliferation and Self Renewal ...... 6

Selective Enrichment of HSCs: Drug-resistance Gene Transfer ...... 13

Dual-gene Transfer Strategies ...... 25

Targeted Transduction ...... 29

Balancing Safety with Efficacy ...... 30

Chapter 2: Viral Vectors for Murine HSCs Gene Therapy Models ...... 33

Dual-gene Vectors for Chronic Granulomatous Disease ...... 34

Lentivector Expression in Hematopoietic Cells ...... 41

Chapter 3: Cotransduction with Separate Single-Gene Lentiviral Vectors ...... 49

Characterization of Cotransduction In vitro ...... 50

Characterization of Cotransduction In vivo ...... 62

Linking Progenitor Selection to Lineage-specific Expression ...... 68

Chapter 4: Comparative Analysis of Dual-Gene Delivery Strategies ...... 83

In Vitro Comparisons ...... 84

i In Vivo Comparisons ...... 95

Chapter 5: Screening Retrovirus Libraries for Mutants with Expanded Host Range ....112

Transduction-based Library Screens ...... 121

Transfection-based Library Screens ...... 124

Chapter 6: Utilizing Phage Peptide Libraries for Targeted Transduction Strategies .....133

Panning Phage Peptide Libraries over SKL Cells In Vitro ...... 134

Panning for SKL-Cell Homing Phage In Vivo ...... 143

Chapter 7: Final Discussion and Future Directions ...... 149

HSC Gene Transfer Vectors ...... 149

Differential Transduction Rates in Murine Progenitor Populations ...... 151

Dual-gene Transfer Strategies ...... 153

Concluding Statement ...... 155

Bibliography ...... 156

ii LIST OF TABLES

Table 1: Summary of In Vitro SKL-Cell Panning Experiments...... 136

Table 2: Summary of In Vivo SKL-Cell Panning Experiments...... 136

Table 3: Phage Clones Isolated from In Vivo SKL-Cell Panning...... 137

Table 4: Phage Clones Isolated from Envelope Panning...... 144

iii LIST OF FIGURES

Figure 1: Enrichment of transduced HSCs in vivo...... 7

Figure 2: Selective amplifier genes (SAGs)...... 10

Figure 3: Drug efflux pumps...... 14

Figure 4: MGMT-mediated repair...... 18

Figure 5: The role of DHFR in pyrimidine biosynthesis...... 24

Figure 6: Strategy for lentiviral cotransduction and selective enrichment...... 28

Figure 7: Bicistronic Retroviral Vectors...... 34

Figure 8: IRES driven MGMT expression levels are reduced...... 35

Figure 9: Lower MGMT expression results in sensitivity to drug treatment...... 36

Figure 10: A human cell line model of X-CGD...... 38

Figure 11: Vector-mediated reconstituted of NADPH oxidase activity...... 39

Figure 12: Dual-gene expression levels after drug treatment...... 40

Figure 13: Promoter strength comparison in hematopoietic cells...... 42

Figure 14: Dual detection of MGMT and GFP by flow cytometry...... 51

Figure 15: Cotransduction and selective expansion levels using a constant MOI...... 53

Figure 16: MGMT and GFP transduction efficiencies are independent...... 54

Figure 17: Cotransduction and selective expansion levels with varying MOIs...... 56

Figure 18: MGMT and dual-gene expressing cells are enriched equally...... 57

Figure 19: Selective expansion of cotranduced murine bone marrow cells...... 59

Figure 20: Selective expansion of cotranduced murine SKL cells...... 61

Figure 21: Cotransduction and in vivo selection of murine BM-MNCs...... 63

Figure 22: Cotransduction and in vivo selection of murine SKL cells...... 65

iv Figure 23: Vector insertions per cell after in vivo selection...... 67

Figure 24: Cotransduction links selection and lineage specific expression in vivo....69

Figure 25: GFP expression in recipient TER119+-BM-MNCs...... 70

Figure 26: Gata-GFP expression is limited to Ter119+ erythroblasts...... 71

Figure 27: Vector copy numbers per cell after in vivo selection...... 72

Figure 28: Enrichment and lineage-specific expression in vivo with limiting MOIs..74

Figure 29: Lentivector constructs for comparison...... 85

Figure 30: Drug resistance levels are equivalent with each strategy...... 86

Figure 31: Comparison of dual-gene transfer strategies in K562 cells...... 87

Figure 32: Relative MFI of MGMT or GFP in K562 cells...... 88

Figure 33: Dual-gene transfer expression levels...... 89

Figure 34: GFP localization after dual-gene transfer...... 89

Figure 35: Dual-gene transfer and expression in murine BM-MNCs...... 91

Figure 36: Relative MGMT or GFP fluorescence in BM-MNCs...... 92

Figure 37: Dual-gene transfer and expression in murine SKL cells...... 94

Figure 38: Relative MGMT or GFP fluorescence in SKL cells...... 95

Figure 39: Dual-Gene expression in PB-MNCs after in vivo selection...... 96

Figure 40: Dual-gene expression in BM-MNCs after in vivo selection...... 99

Figure 41: Dual-gene expression in lymphoid and myeloid populations...... 99

Figure 42: Vector insertion averages are equivalent with each strategy...... 101

Figure 43: Relative MGMT and GFP MFI in BM-MNCs...... 102

Figure 44: Structural alignment of MuLV envelope receptor-binding domains...... 115

Figure 45: The ecotropic envelope receptor binding domain...... 116

v Figure 46: Randomization of the ecotropic envelope gene for retroviral display.....117

Figure 47: Production of retroviral envelope libraries...... 118

Figure 48: Viral tropism and receptor interference in packaging cell lines...... 119

Figure 49: Env-expressing retroviral vectors are amplified in gag+-pol+ cells...... 120

Figure 50: Exponential transfer of an Env-expressing RV vector in gag+-pol+

cells...... 121

Figure 51: Transduction-based screens in E86 cells...... 122

Figure 52: Transduction-based screens in 293 cells...... 123

Figure 53: Transduction-based screens in AM12 cells...... 124

Figure 54: Transfection-based screens in AM12 cells...... 125

Figure 55: Transfection-based screens in E86 cells...... 126

Figure 56: Transfection-based screens in TEFLY cells...... 127

Figure 57: Panning phage libraries for stem cell binding epitopes...... 135

Figure 58: Phage clone binding to murine BM-MNCs...... 138

Figure 59: Phage BM-MNC binding in the presence of unlabeled competitor...... 139

Figure 60: Cyclic peptides synthesized...... 140

Figure 61: Titration of the SKL peptide on murine PB and BM MNC fractions...... 141

Figure 62: Murine BM-MNC fractions bound by cyclic peptides...... 142

vi PREFACE

This dissertation describes novel research aimed at improving dual-gene transfer

and vector targeting for hematopoietic stem cell gene therapy. HSC gene therapy is

limited by the low numbers of gene-corrected cells that can be attained in vivo; due to low HSC gene transfer efficiency and the low proportion of gene-corrected cells after

transplant. The first chapter of this dissertation reviews the current gene therapy vectors

and methods for enriching transduced HSCs. Previous studies in our lab have

demonstrated that small populations of HSCs transduced with the drug-resistance gene,

MGMT, can be selectively enriched in vivo with drug selection. The second chapter

describes research aimed at linking this selection strategy to therapeutic gene expression

in order to reconstitute diseased hematopoietic tissue with gene-corrected cells. This

strategy was evaluated using a murine model of X-linked chronic granulomatous disease.

Although stem cell enrichment was achieved, only a low level of therapeutic gene

activity was detected in vivo. Thus, the research emphasis was shifted away from

specific gene therapy models towards optimizing dual-gene transfer and expression in

murine HSCs. Several retroviral vectors and promoters were evaluated for optimal gene

transfer and expression in murine HSCs. Limitations arising from traditional dual-gene

vectors prompted investigations into using separate single-gene vectors, including

MGMT, to cotransduce cells and selectively amplify dual-gene expressing populations.

The third chapter presents a detailed analysis of the vector MOI ratios resulting in the

greatest enrichment of dual-gene expressing cells. Further studies demonstrated that this

strategy could be used to couple MGMT-mediated progenitor cell selection to expression

of another, lineage-specific gene; an essential requirement for many gene therapy

vii applications. The fourth chapter describes comparative studies between cotransduction with single-gene vectors and transduction with traditional dual-gene vectors, with dual- gene expressing cell enrichment, expression levels, and vector insertion averages as the basis for comparisons. Cotransduction with single-gene vectors was found to be an efficient alternative to the dual-gene vectors and was less prone to complications that affected protein expression or processing.

Another limitation in HSC gene therapy is the inability to specifically transduce stem cells in mixed donor cell populations. Several strategies have focused on engineering known ligands into viral envelope proteins to target the virions to specific cell types. However, rational engineering approaches have had little success, due to unpredicted perturbations in envelope function. We have initiated a series of combinatorial approaches to this problem. The first strategy, presented in chapter five, involved randomizing the receptor-binding domain of the ecotropic virus envelope protein, using a unique screen to identify mutations that conferred unconventional cell entry mechanisms. Another strategy, described in chapter six, involved panning stem cells with phage peptide libraries to pare down diverse library sizes to populations that are both biased for the target cells and fit within the constraints of eukaryotic display methodologies. The significance of these data, and suggestions for future experiments, are discussed in chapter seven.

viii ACKNOWLEDGEMENTS

I would like to thank Stan Gerson for the guidance and wealth of opportunities he

provided, and for his flexibility and patience in letting me pursue some of my "hair-

brained" ideas.

I would like to thank my committee members for their time and careful

consideration of my studies; particularly, Ed Stavnezer and Eric Arts for their vital insight and continual attempts at keeping me on track, and Kevin Bunting and Rolf

Renne for thoughtful discussions above and beyond the call of duty.

For their assistance, and for providing a most enjoyable work environment, I would like to extend my appreciation to past and present members of the Gerson lab;

Anthony Benson, Chris Ballas, Colin Sweeney, Jane Reese, Jon Donze, Karen Lingas,

Lili Liu, Min Liu, Mourad Ismail, Steve Zielske, Yuan Lin, and Youngji Park. Special thanks go to Mourad Ismail and Karen Lingas for their ability to pick up and run with an experiment at a moments notice.

I would also like to thank Kate Reinicke for her persistence in ensuring I have a life outside of the lab.

This research was supported by the Public Health Service grants, R01CA073062, and P30CA43703, and the T32CA059366 and T32HL07147 training grants.

ix LIST OF ABBREVIATIONS

5-FU: 5-fluorouracil

AGT: O6-alkylguanine-DNA alkyltransferase

AM12: murine amphotropic packaging cells

Ampho: Amphotropic

APC: Allophycocyanin

BM: bone marrow

BM-MNC: bone marrow mononucleocyte

CAT: cationic amino acid transporter

CDA: cytidine deaminase

CFU: colony-forming unit

CGD: chronic granulomatous disease

cGy: centi-Gray

CID: chemical inducer of dimerization

CMV: cytomegalovirus

cPPT: central polypurine tract

CTS: central termination sequence

DHFR: dihydrofolate reductase

DMEM: Dulbecco’s modified Eagle medium

E86: murine ecotropic packaging cells

Eco: Ecotropic

EF-1α: elongation factor-1α

EIG: Envelope-IRES-GFP

x EMCV: encephalomyocarditis virus env: envelope

EPO: erythropoietin

FBS: fetal bovine serum

FMDV: foot and mouth disease virus

GALVR: gibbon ape leukemia virus receptor

GAPDL4: glyceraldehyde-3-phosphate dehydrogenase-like 4

GFP: green fluorescence protein

GP: gag-pol gp91phox: X-linked surface subunit of NADPH oxidase

HIV: human immunodeficiency virus

Hox: homeobox

HSC: hematopoietic stem cell

IL: interleukin

IMDM: Iscove's Modified Dulbecco's Medium

IRES: internal ribosome entry site

K562: human erythroleukemia cell line

LV: lentivirus

MAG: MGMT(P140K)-FMDV2a-GFP mdr1: multidrug-resistance-1

MEM: minimum essential medium

MFI: mean fluorescence intensity

MGMT: O6-methylguanine-DNA methyltransferase

xi MIG: MGMT(P140K)-EMCV-IRES-GFP

MIP: MGMT-EMCV-IRES-gp91phox

MMR: mismatch repair

MND: MPSV promoter, deleted negative control region, substituted primer binding site

MOI: multiplicity of infection

MPSV: myeloproliferative sarcoma virus

MSCV: Murine Stem Cell Virus retroviral vector

MTX: methotrexate

MuLV: murine leukemia virus

NADPH: nicotinamide adenine dinucleotide phosphate, reduced form

NBMPR-P: nitrobenzylmercaptopurineriboside phosphate

NBT: nitroblue tetrazolium

NOD: non-obese diabetic

P140K: Proline-140 substitution with lysine

PAM: gp91phox-FMDV2a-MGMT

PB-MNC: peripheral blood mononucleocyte

PBS: phosphate-buffered saline

PCR: polymerase chain reaction

PE: Phycoerythrin

PEG: polyethylene glycol

PFU: plaque-forming units

PGK: phosphoglycerate kinase

P-gp: P-glycoprotein

xii phage: bacteriophage phox: phagocytic oxidase, NADPH oxidase

PIM: gp91phox-EMCV-IRES-MGMT(P140K)

PMA: phorbol ester 12-tetradecanoylphorbol-13 acetate

PRE: post-transcriptional regulatory element

RBD: receptor binding domain

RD114: feline endogenous virus envelope protein rre: rev response element

RV: retrovirus

SAG: selective amplifier gene

SCF: stem cell factor

SCID: severe combined immuno-deficiency

SIN: Self-inactivating

SKL: Sca+/Kit+/Lineageneg

SU: surface subunit

TEFLY: human fibrosarcoma GP-cell line

TM: transmembrane domain

VR: variable region

VSVG: vesicular stomatitis virus G glycoprotein

X-CGD: X-linked chronic granulomatous disease

X-SCID: X-linked severe combined immuno-deficiency

xiii Dual-Gene Transfer and Vector Targeting for Hematopoietic Stem Cell Gene Therapy

Abstract

by

JUSTIN CHARLES ROTH

Hematopoietic stem cell gene therapy is limited by ex vivo gene transfer efficiency and by the number of gene-corrected cells that can be attained in vivo.

Although lentiviral vectors have improved HSC gene transfer rates, cytokine stimulation

and high vector MOIs are still required for efficient transduction of these cells. Our lab

has demonstrated that limiting numbers of HSCs transduced with point mutants of the

DNA repair gene, MGMT, can be efficiently enriched in vivo with BG and BCNU

treatment. Therefore, linkage of MGMT and therapeutic gene expression offers the

potential for reconstituting diseased hematopoietic tissue with gene-corrected cells.

Towards this endpoint, different retroviral and lentiviral vectors were evaluated for

efficient gene transfer and expression in murine HSCs. Traditional dual-gene vector

strategies that directly coupled MGMT to a therapeutic gene were limited by low drug

resistance or poor therapeutic gene expression. Therefore, a new strategy was devised;

using separate single-gene lentivirus vectors, including MGMT, to cotransduce cells and

enrich dual-gene expressing populations. The stoichiometry of the two single-gene

vectors and the total vector MOIs required for efficient cotransduction and dual-gene

expressing cell enrichment are described. This information was then used to cotransduce

xiv and selectively enrich dual-gene expressing murine bone marrow cells in vivo. Murine transplants were also carried out using donor cells cotransduced with a ubiquitous

MGMT lentivector and an erythroid-specific GFP vector. These studies showed the potential of cotransduction for obtaining therapeutic gene expression in lineage-specific progeny of MGMT-enriched progenitor cells. Dual-gene vectors containing EMCV-

IRES or FMDV-2a elements for co-expression were then compared to the single-gene vector cotransduction strategy. These data demonstrated that cotransduction is an efficient alternative to dual-gene vectors and offers flexibility in application, since it avoids the need to construct dual-gene vectors and establish drug selection conditions for

each new therapeutic application. Additional studies were initiated with the aim of

specifically targeting virions to stem cells. Novel approaches towards the development of

targeted virions are described. These reagents should reduce the risk of insertional

mutagenesis by decreasing integration events in non-target cells, and the MOIs required

for efficient HSC transduction in mixed cell populations.

xv CHAPTER 1

HEMATOPOIETIC STEM CELL GENE THERAPY

*Portions of this chapter are from:

Justin C Roth and Stanton L Gerson. "Drug Resistance Gene Transfer to Hematopoietic Stem Cells," In: Gene Therapy for Cancer; Humana Press, In press.

Introduction

Once achieved, efficient and stable gene transfer into hematopoietic stem cells

(HSCs) has the potential for curing hematologic disease, or alleviating the myelosuppressive consequences of antineoplastic chemotherapy. Regardless of the therapeutic endpoint, HSC gene transfer efficiency is limited, and the proportion of gene- modified cells is further diluted by the vast excess of unmanipulated cells upon transplant. These limitations have spurred efforts aimed at improving the specificity of

HSC gene therapy vectors and the development of methods for enriching gene-modified stem cells in vivo. This chapter will describe the common viral vector systems used for stable gene transfer into HSCs, and the main strategies used for protecting and enriching these cells in vivo.

HSC GENE THERAPY

Stable Gene Transfer Vectors

The requirement for high gene transfer and expression levels in HSC gene therapy has been a driving force in the development of viral vector delivery systems. Viruses that integrate into the host genome, such as adeno-associated viruses (AAV), retroviruses

(RV), and lentiviruses (LV), are of particular interest due to their stable transmission and

1 expression of the transgene. Significant progress has been made in identifying cis- and

trans-acting elements, from endogenous and exogenous sources, that influence the

efficiency of vector packaging, integration, and expression.

Vectors derived from murine leukemia virus (MuLV), myeloproliferative sarcoma virus (MPSV), spleen focus-forming virus (SFFV), and other oncoretroviruses have been generated (Artelt et al. 1988; Hawley et al. 1992; Tumas et al. 1996; Halene et al. 1999).

Extensive characterization of these vectors has revealed both positive and negative regulatory elements that affect transgene expression in primitive hematopoietic cells

(Baum et al. 1999; Halene et al. 1999; Cherry et al. 2000; Ketteler et al. 2002).

Exogenous sequences have also been inserted into these vectors to improve transgene expression levels. Scaffold attachment regions (Agarwal et al. 1998; Murray et al. 2000) and insulator elements (Inoue et al. 1999; Emery et al. 2000; Rivella et al. 2000) reduce silencing and positional effects, while post-transcriptional regulatory elements (PREs), such as that from the Woodchuck hepatitis virus, improve expression levels by increasing transcript export and stability (Popa et al. 2002). The cell division requirement for retroviral integration and the quiescent nature of HSCs has limited retrovirus-based transduction efficiencies. However, insight into the roles specific cytokines play in hematopoiesis has led to the identification of cytokine cocktails that promote stem cell cycling and transduction, while limiting commitment towards differentiation (Dao et al.

1997; Luens et al. 1998).

Many of the recent HSC gene transfer models have utilized lentivirus vectors, such as those based on the human immunodeficiency virus (HIV-1). Lentiviruses actively transport the reverse-transcribed transgene across the nuclear envelope and

2 therefore do not require cell division for access to the host genome (Naldini et al. 1996).

This active-transport enables lentiviruses to transduce non-dividing cells (Reiser et al.

1996). However, lentiviral transduction of HSCs are dramatically improved after cytokine stimulation, indicating that at least partial entry into the cell cycle may be necessary for key steps in viral entry (Sutton et al. 1999; Douglas et al. 2001; Zielske et al. 2003). As with retroviral vectors, key cis-acting elements in the lentiviral backbone have been identified as essential components for efficient transduction and expression.

The central DNA flap, composed of the central polypurine tract and the central termination sequence (cPPT/CTS), has been shown to be an important element for nuclear translocation of the reverse-transcribed genome (Zennou et al. 2000). Re- insertion of this sequence into lentiviral vectors was shown to dramatically improve transduction of HSCs (Follenzi et al. 2000; Sirven et al. 2000). Self-inactivating (SIN) vectors, originally described for retroviral vectors and duplicated in lentiviral vectors, have been generated by deleting promoter and enhancer sequences from the U3 region of the 3’ LTR (Yu et al. 1986; Miyoshi et al. 1998; Zufferey et al. 1998). Following reverse transcription, this modification is duplicated to form the 5' U3 region. Thus, promoter and enhancer functions are effectively removed from both proviral LTRs. SIN vectors are a major improvement in safety, as they reduce the capacity for recombination or rescue from exogenous virus, and deletion of the strong LTR enhancer sequences may reduce the frequency of insertional activation. Transgene expression in SIN vectors is achieved by the insertion of an internal promoter, a range of which have been evaluated for HSC gene therapy, including the ubiquitous cytomegalovirus (CMV), elongation factor-1α (EF-1α), and phosphoglycerate kinase (PGK) promoters (Sutton et al. 1999;

3 Salmon et al. 2000). Vector comparisons utilizing the GFP reporter indicate that the EF-

1α promoter provides the most robust multilineage hematopoietic expression levels

(Salmon et al. 2000). Additional studies have utilized cellular promoter and enhancer sequences to restrict expression to specific hematopoietic lineages (Moreau-Gaudry et al.

2001; Lotti et al. 2002). Many of the enhancer sequences developed in oncoretroviral vectors for strong hematopoietic expression have also been functionally transferred into the lentivector systems (Choi et al. 2001; Demaison et al. 2002). The first clinical trial using lentiviral vectors was recently approved for anti-HIV therapy (Manilla et al. 2005).

This particular study should also determine whether or not SIN lentivectors are truly innocuous in the presence of high loads of wild-type virions.

Envelope Pseudotypes

Viral envelope proteins are responsible for cell type specificity and viral membrane stability. In addition to combining sequence elements for improved vector expression, vast efforts have gone into evaluating envelope proteins from other viral species for functional substitution of retro- and lentivirus envelope proteins to improve virion stability and transduction efficiency. This type of substitution is known as pseudotyping. Envelope proteins can often be pseudotyped without prior modification but higher expression levels may be required to overcome lost envelope-capsid interactions. The amphotropic murine and gibbon ape leukemia virus envelope proteins were favored in early hematopoietic gene transfer applications, but subsequent studies have shown that the receptors for these envelopes are limiting in the more primitive stem

4 cell populations (Thomsen et al. 1998). The reduced stability of these envelope proteins

also limits production of concentrated viral stocks.

The vesicular stomatitis virus G glycoprotein (VSVG) is the most widely used

pseudotype for lentivirus gene transfer. This protein allows the transduction of a wide

range of species and cell types, as it utilizes a ubiquitous phospholipid for membrane

fusion, rather than a protein receptor, for entry (Carneiro et al. 2002). In addition,

VSVG-pseudotyped viruses are much more stable, allowing virus-enriched media to be

concentrated to levels over 108 transducing units/ml. The main drawback of VSVG is its cytotoxicity. This toxicity has prompted the development of inducible VSVG packaging cell lines (Farson et al. 2001; Pacchia et al. 2001). Despite the reduced expression of the amphotropic receptor in early stem cells, von Laer et al. reported that retroviral particles pseudotyped with VSVG or the amphotropic envelope appear to transduce human hematopoietic progenitors with similar efficiency (von Laer et al. 2000).

Vectors pseudotyped with the feline endogenous virus envelope protein, RD114, have also been described (Kelly et al. 2001). RD114-pseudoyped virions can be concentrated and the lack of toxicity has permitted the generation of stable packaging cell lines (Ward et al. 2003). Hanawa et al. compared lentivectors pseudotyped with VSVG,

RD114, or amphotropic envelopes for their ability to be concentrated and transduce primitive human hematopoietic CD34+ cells (Hanawa et al. 2002). The amphotropic and

RD114 particles excelled over VSVG at transducing cord blood-derived CD34+

progenitors, and the amphotropic particles were more efficient at transducing SCID

repopulating cells derived from G-CSF-mobilized peripheral blood CD34+ cells.

Relander et al. recently compared lentivectors pseudotyped with modified gibbon ape

5 leukemia virus, RD114, or amphotropic envelopes for transduction efficiency of

NOD/SCID-repopulating human CD34+ cells (Relander et al. 2005). In the presence of cytokine stimulation, the modified RD114-pseudotyped vectors generated the highest percentage of transduced cells. Many other pseudotypes are being evaluated for use in a variety of applications. The best pseudotype for any particular situation will ultimately depend on several factors, including the type of vector, the transduction protocol, and the

type and source of the target cells.

HSC EXPANSION: INDUCERS OF PROLIFERATION AND SELF-RENEWAL

Limited gene transfer efficiency into HSCs and the myelotoxicity associated with

neoplastic drug treatments have led to efforts aimed at enriching gene-modified cells in

vivo. Genes that are utilized to specifically enrich transduced stem cells in vivo can be

divided into two main classes: those that induce HSC proliferation and expansion, and

those that protect cells from cytotoxic drug treatments (Fig. 1). The particular strategy

employed will likely be determined by the type and severity of the disorder for which

gene transfer is needed. Disorders such as chronic granulomatous disease that can be

alleviated by low levels of gene-corrected cells may favor stem cell expansion strategies.

Other diseases, such as sickle cell anemia in which uncorrected cells negatively impact

on survival, may favor drug selection strategies. Multi-gene vectors offer the additional

potential for linking the advantages of each strategy and will be discussed later in the

chapter.

6

A. B.

Figure 1: Enrichment of transduced HSCs in vivo. Transduced HSCs can be enriched in vivo using genes that provide a proliferative advantage (A) or a survival advantage (B) over untransduced cells.

7 Endogenous Genes for HSC Expansion

Gene products that contribute to stem cell self renewal and proliferation are

highly sought after as tools for overcoming the limited cell numbers associated with stem cell therapies. Considerable progress has been made in identifying growth factors that can maintain or expand the stem cell pool ex vivo (Sauvageau et al. 2004). However,

growth factors act on natural receptors, and thus cannot be used to specifically enrich

transduced stem cell population in vivo. Therefore, genes that provide an intrinsic

proliferative advantage to stem cells are being evaluated as tools for enriching transduced

populations. Endogenous genes, which have a natural role in stem cell self renewal, as

well as recombinant genes engineered for this activity, have been assessed as candidates

for this approach.

Although gene products that have a natural role in stem cell self renewal and

proliferation will likely have the most potential for this type of application, many have

also been implicated in leukemogenesis. The homeobox transcription factors are a good

example of this. This gene family was initially discovered for its role in embryogenesis,

but many of the hox genes also participate in hematopoiesis. Several members from the

hoxA, B, and C gene clusters have distinct expression patterns that are restricted to

different lineages or stages of hematopoietic differentiation (Giampaolo et al. 1994;

Moretti et al. 1994; Sauvageau et al. 1994). Ectopic expression of these genes, either

with retroviral vectors, or by naturally occurring translocation events, has also been

linked to major perturbations in hematopoiesis (Grier et al. 2005). The HoxB4

transcription factor appears to be an exception. Transduction of murine or human HSCs

with HoxB4 has been shown to induce expansion in vitro and in vivo (Sauvageau et al.

8 1995; Thorsteinsdottir et al. 1999; Antonchuk et al. 2002; Amsellem et al. 2003; Krosl et

al. 2003; Schiedlmeier et al. 2003). Murine HSCs overexpressing HoxB4 have up to a

50-fold competitive repopulation advantage over untransduced cells. In these HoxB4 studies stem cell expansion did not progress beyond normal stem cell levels, which suggests the existence of an environmental sensor of stem cell density (Thorsteinsdottir et al. 1999). The level of HoxB4 expression also seems to determine its biological activity.

Beslu et al. demonstrated that increased HoxB4 expression levels correlated with increased repopulating potential (Beslu et al. 2004). HoxB4 applied to cells in protein form has also been shown to allow ex vivo stem cell expansion (Amsellem et al. 2003;

Krosl et al. 2003). This strategy offers the potential for pre-loading stem cells prior to

transplant, allowing transient expansion without a requirement for stable expression.

Delivery of other hox members with this strategy may allow for transient expansion of

specific lineages without the associated risk of leukemogenesis. HoxB4 expression in

human CD34+ cord blood cells has been reported to impair lymphomyeloid

differentiation (Schiedlmeier et al. 2003). Thus, the necessary feedback signals that

appear to be present in the all-murine model may not be recognized by human cells when

expanded ex vivo or transplanted into NOD/SCID mice.

Recombinant Genes for HSC Expansion

Another strategy for stem cell expansion utilizes chimeric gene products, called

selective amplifier genes (SAGs). SAGs are composed of a dimerization and a signaling

domain, which become activated by specific molecules called chemical inducers of dimerization (CID). Dimerization activates the SAG signaling domains to induce cell

9 proliferation (Fig. 2). Several variations of this theme have been evaluated for both in

vitro and in vivo expansion in response to CID administration.

Figure 2: Selective amplifier genes (SAGs). SAGs are chimeric gene products consisting of dimerization and signaling domains. Chemical inducers of dimerization (CIDs) bring two SAG monomers together, which activates the signaling domains to transmit a cell proliferation signal.

Initial studies utilized the FK506-binding domain of the immunophilin FKBP12

linked to the intracellular signaling domain of either the erythropoietin or c-kit receptors

(Blau et al. 1997; Jin et al. 1998). The feasibility of this strategy was demonstrated using

an IL3-dependent cell line; addition of the CID (FK1012) to the media rescued

transduced cells from IL3 deprivation (Blau et al. 1997; Jin et al. 1998). The intracellular

10 signaling domain of the thrombopoitin receptor was evaluated in subsequent studies (Jin

et al. 1998). Murine bone marrow cells transduced with this construct could be expanded

ex vivo only in the presence of FK1012. Although multilineage expansion was demonstrated at early time points, megakaryocytic cells dominated the cultures at later time points. CID-mediated expansion of CD34+ cord blood progenitors was achieved in

later experiments using a similar construct (Richard et al. 2000). Whereas the expanded

murine cultures favored megakaryocytic differentiation, erythroid cells dominated the

CID-expanded human cell cultures.

SAG-mediated HSC expansion has also been evaluated in vivo. Zhao et al.

recently demonstrated that SAGs derived from Jak family members may be useful for

amplifying specific hematopoietic lineages (Zhao et al. 2004). Experiments were carried

out with a SAG construct containing the JH1 domain of murine Jak2 linked to a tandem

binding site for the CID, AP20187, and evaluated in a murine transplant model.

Administration of the CID resulted in a rapid expansion of transduced erythrocytes.

However, the effect was short-lived, and the transduced erythrocyte population declined to pretreatment levels after CID withdrawal. Another SAG, consisting of the erythropoietin (EPO) receptor dimerization domain and the thrombopoitin receptor signaling domain, was recently evaluated in cynomolgus macaques (Ueda et al. 2004).

Transduced CD34+ cells were transplanted directly into irrigated femurs and humeri in

unconditioned animals. In the absence of CID administration (in this case EPO), 2-30%

of colony-forming unit (CFU) cells and less than 0.1% of peripheral mononucleocytes were transgene positive after 1 year. Peripheral blood marking levels in animals treated with daily injections of EPO peaked at 8-9% over the same time period, with polyclonal

11 marking detected in multiple lineages. However, as seen in previous studies, marked cell

percentages returned to baseline levels shortly after each CID treatment (Ueda et al.

2004).

Selective amplification strategies reported to date have demonstrated promising

results. However, the marking levels obtained in animal models remain low, and depend

on continuous CID administration. It remains unclear whether the chimeric gene

products are only effective at expanding less primitive cell populations, or decline as a result of an immune response. However, rather than returning to baseline levels, an immune response would likely clear all SAG-expressing cells upon CID withdrawal.

One advantage SAG-mediated stem cell expansion has over drug selection schemes is the limited toxicities associated with selection. No obvious adverse events were detected from transgene expression or CID treatment in the animal models. Nevertheless, no attempts have been made to determine whether cells transduced with these constructs exhibit normal cell cycle checkpoint controls in the setting of such strong proliferative signals. This issue needs to be addressed, especially if attempts will be made to link SAG and drug selection strategies. Transcriptional profiling of stem cells should identify additional candidates for this approach.

12 SELECTIVE ENRICHMENT OF HSCs:

DRUG-RESISTANCE GENE TRANSFER

Drug resistance genes offer another approach to enriching transduced stem cell

populations in vivo. Whereas selective amplifier genes confer a proliferative advantage,

allowing transduced stem cells to outgrow untransduced cells, drug resistance genes

provide a survival advantage to transduced stem cells in response to cytotoxic drug

treatment. The most common restriction to neoplastic drug treatment is the associated myelotoxicity. Chemotherapy resistant genes transferred to HSCs should reduce treatment-related morbidity and permit dose escalation to levels needed for neoplastic cell toxicity. In addition, drug resistance genes offer the potential to specifically select transduced cells at the expense of unmodified cells in vivo. Several drug resistance genes have been evaluated as tools for either increasing the therapeutic index of cancer therapies, or for selective enrichment of gene corrected stem cells.

Multidrug-Resistance Proteins

P-glycoprotein (P-gp) is encoded by the multidrug-resistance-1 gene (mdr1) and is the prototypic member of the ATP-binding cassette family of drug resistance proteins.

P-gp was identified as the cellular protein responsible for the pleiotropic cross-resistance certain cell lines acquired to unrelated chemotherapy drugs, such as anthracyclins, epipodophyllotoxins, taxanes, and vinca-alkaloids (Juliano et al. 1976; Kartner et al.

1985). Drug resistance is established by the ATP-dependent efflux of these compounds from the cell, preventing intracellular concentrations from becoming cytotoxic (Fig. 3).

13 Figure 3: Drug efflux pumps. The ABC-family of drug efflux proteins are localized in the plasma membrane where they actively export cytotoxic drugs from the cytosol to the extracellular space. Cells that overexpress these proteins are protected from higher drug concentrations.

The potential for using P-gp to protect bone marrow from myelosuppression was first demonstrated in transgenic mice (Galski et al. 1989; Mickisch et al. 1991). High level P-gp expression in transgenic bone marrow protected animals from daunomycin and taxol treatments that caused myelosuppression in normal mice. Transplantation of mdr1- transgenic bone marrow into lethally irradiated control animals was sufficient to transfer

14 long-term drug resistance to recipient animals (Mickisch et al. 1992). These results set

the framework for retroviral gene transfer experiments. In 1992 two groups

demonstrated that retroviral delivery of mdr1 to bone marrow progenitor cells resulted in

increased drug tolerance that correlated with enrichment for the transduced population

(Podda et al. 1992; Sorrentino et al. 1992). Transduction of long-term repopulating cells

with mdr1 was subsequently shown to protect murine transplant recipients from repetitive

administration of normally myelotoxic chemotherapy treatments (Hanania et al. 1994;

Carpinteiro et al. 2002), while simultaneously sensitizing tumor cells (Hanania et al.

1997). More recent in vivo selection studies using mdr1 have been carried out in other animal models. Schiedlmeier et al. demonstrated the first significant evidence of mdr1- mediated selection of human hematopoietic progenitors in vivo, using NOD/SCID recipients treated with paclitaxel (Schiedlmeier et al. 2002).

Clinical trials have also been carried out to evaluate the use of mdr1 gene transfer for cancer patients receiving autologous transplants to reduce treatment-induced myelosuppression (Hanania et al. 1996; Hesdorffer et al. 1998; Vahdat et al. 1998;

Cowan et al. 1999; Moscow et al. 1999; Abonour et al. 2000). Early trials were limited

by poor transduction efficiencies and resulted in transient marking levels indicative of

short-term repopulating cell contributions. Higher marking rates (up to 52%) have been

obtained using plates coated with the recombinant fibronectin fragment CH-296

(Abonour et al. 2000). A maximum of 15% mdr1-marked bone marrow CFU were

detected a year after transplant, but evidence of mdr1-mediated selection was limiting.

More than 50 other ATP-binding cassette family members have been identified, some of which have been associated with antineoplastic drug resistance in cancer cells

15 (Abbott 2003). In experiments aimed at resolving the cycling status of HSCs, Goodell et al. discovered a population of cells resistant to labeling with the fluorescent DNA stain,

Hoechst 33342 (Goodell et al. 1996). Interestingly, lymphomyeloid repopulating activity was enriched over 1000 fold in the population of cells with the highest degree of Hoescht exclusion (designated SP cells). The verapamil-sensitivity of this phenomenon indicated that P-gp, or a similar drug efflux pump, was responsible and expressed at higher levels in bone marrow stem cells. Subsequent studies have identified SP stem cells in a variety of tissues. ABCG2/BCRP1 has been identified as the transporter responsible for Hoechst

33342 efflux in SP cells. However, enforced ABCG2 expression in murine bone marrow reduced progenitor differentiation in vitro (Zhou et al. 2001; Zhou et al. 2002). Similar

results were previously reported for murine bone marrow cells transduced with mdr1

(Bunting et al. 2000). Overexpression of mdr1 in the absence of drug selection resulted

in SP cell expansion ex vivo and the onset of a myeloproliferative syndrome when these

cells were transplanted into lethally irradiated recipients. Another group has recently

reported an increased frequency of leukemogenesis associated with retroviral delivery of

mdr1 to murine BM cells at a high copy number (Modlich et al. 2005). However, in this

study insertional mutagenesis was also seen using fluorescent reporter vectors, but at a reduced frequency. The high expression levels of these proteins in primitive hematopoietic cells suggest that these transporters may have a role in stem cell biology beyond protection from drug exposure (Kim et al. 2002; Scharenberg et al. 2002).

However, both mdr1 and ABCG2 knockout animals exhibit normal hematopoiesis

(Uchida et al. 2002; Zhou et al. 2002). No myeloproliferative disorders were detected during the mdr1 clinical trials. Further, no aberrant expansion was detected in rhesus

16 macaques after transplanting cells transduced with conditions similar to those that caused

the disorder in mouse models (Sellers et al. 2001).

Drug selection of stem cells transduced with mdr1 or ABCG2 may be limited by the high endogenous levels of these proteins in primitive hematopoietic cell populations.

Specific point mutations of mdr1 (Hafkemeyer et al. 2000) or ABCG2 (Ujhelly et al.

2003) that are resistant to inhibition, or have altered substrate recognition, may be more

potent agents for differentially selecting transduced stem cells in vivo.

Alkylating Agent Resistance

The most striking HSC selection results in vivo have been obtained using the O6- methylguanine-DNA methyltransferase (MGMT) gene. MGMT encodes O6-

alkylguanine-DNA alkyltransferase (AGT), which repairs DNA damage induced by

alkylating agents (Fig. 4). Although most DNA repair pathways involve multiple protein

constituents, AGT is singly responsible for the repair of O6-alkyl lesions. Repair is

mediated by the covalent transfer of the O6-alkyl group from guanine to a cysteine thiol

located in the AGT binding pocket (Pegg et al. 1983). This irreversible reaction

inactivates AGT. Thus, each AGT molecule is only capable of repairing one alkyl lesion,

after which the protein is ubiquitinated and targeted for degradation (Srivenugopal et al.

1996).

17 Figure 4: MGMT-mediated repair. MGMT repairs cytotoxic O6-alkylguanine lesions formed by methylating and chloroethylating agents. BG inactivates endogenous MGMT, thereby increasing the sensitivity of untransduced cells to alkylating agent treatment. Specific MGMT point mutants (MGMT’) are resistant to BG inactivation, but maintain the capacity for DNA repair. HSCs transduced with MGMT’ are enriched in vivo following BG and alkylating agent treatment.

Alkylation of the O6 position of guanine is the most cytotoxic lesion produced by

methylating (e.g. temozolomide, streptozotocin, dacarbazine, and procarbazine), and

chloroethylating (e.g. BCNU, and CCNU) agents used to treat a variety of cancers.

These agents are particularly myelosuppressive due to the low level of AGT activity in

bone marrow cells (Gerson et al. 1985; Gerson et al. 1987). During DNA replication O6-

methylguanine residues are mismatched with thymine (Moriwaki et al. 1991). Therefore,

DNA synthesis prior to AGT-mediated repair results in a G:T mispair that is corrected by

18 the mismatch repair (MMR) pathway. Uncorrected methylguanine residues result in a

futile MMR cycle in which thymine residues are continuously mispaired opposite O6-

methylguanine, eventually leading to single strand breaks and cell death (Karran et al.

1994). The types of DNA damage induced by chloroethylating agents are particularly

cytotoxic. Unrepaired O6-chloroethyl lesions rearrange to form both intra- and

interstrand crosslinks to neighboring residues (Dolan et al. 1988; Gonzaga et al. 1989).

The significance of using drug resistance gene transfer to protect mammalian cells

from DNA damage was first established after the bacterial MGMT homologue (ada) was cloned (Margison et al. 1985). Transfer and expression of ada in MGMT deficient cell lines was shown to dramatically reduce alkylating agent toxicity (Brennand et al. 1986;

Ishizaki et al. 1986; Samson et al. 1986). These pivotal experiments set the stage for the last two decades of research aimed at using MGMT gene transfer to protect bone marrow cells from the myelosuppressive effects of alkylating agent chemotherapy, and as a mechanism for selecting transduced stem cells in vivo.

The first study to demonstrate MGMT-mediated protection of bone marrow (BM) cells was carried out using electroporation for ada gene delivery (Jelinek et al. 1988).

Stable transfer of ada (Harris et al. 1995) or human MGMT (Allay et al. 1995; Moritz et al. 1995) into murine BM cells with retroviral vectors was subsequently shown to reduce

the myelosuppressive effects of chloroethylating agents in vivo. Increased resistance to

multiple doses of BCNU correlated with increased percentages of MGMT-transduced

murine progenitors in the bone morrow (Allay et al. 1997). Human CD34+ hematopoietic

progenitor cells transduced with MGMT were also shown to tolerate higher doses of

BCNU (Allay et al. 1996).

19 Just as high BM AGT expression reduces alkylating agent-induced myelosuppression, tumor cells with upregulated AGT are also tolerant to these treatments

(Brent et al. 1985; Schold et al. 1989; Citron et al. 1991). Therefore, the modest levels of

protection achieved by MGMT gene transfer experiments are unlikely to have a dramatic

therapeutic impact. However, two major advancements brought MGMT-mediated

chemoprotection to the forefront. First was the discovery of the potent MGMT

inactivator, O6-benzylguanine (BG). BG provided a mechanism for depleting AGT

activity, sensitizing tumors to drug treatment (Dolan et al. 1990; Dolan et al. 1993; Pegg

et al. 1995). However, BG-mediated inactivation of AGT is not specific, and thus

sensitizes both tumor and bone marrow cells to alkylating agents (Fairbairn et al. 1995).

The second major advancement came from the identification of specific point mutations

in MGMT that conferred significant resistance to BG inactivation without altering the O6- alkyltransferase activity (Crone et al. 1994). Additional BG-resistant MGMT mutants were identified from randomized MGMT libraries using BG and O6-alkylating agent

selection schemes (Christians et al. 1997; Xu-Welliver et al. 1998). Specific MGMT

mutants were then shown to efficiently protect transduced human bone marrow

progenitors from BG-mediated sensitization to chloroethylating (Reese et al. 1996) and

methylating (Hickson et al. 1998) agent toxicity (Fig. 4). Davis et al. demonstrated that murine bone marrow progenitors transduced with the BG-resistant MGMT-G156A point mutant could be enriched in vivo with combined doses of BG and BCNU, and this enrichment protected transplant recipients from doses that were lethal to animals transplanted with control bone marrow cells (Davis et al. 1997). Other MGMT point

mutants were also shown to protect mice from combined BG and temozolomide

20 (Chinnasamy et al. 1998), or BG and BCNU treatments (Ragg et al. 2000). These studies

also demonstrated that selective enrichment occurred at the stem cell level.

The true potential for using MGMT mutants for stem cell selection was

demonstrated by Davis et al., using nonmyeloablated transplant recipients (Davis et al.

2000). Transduced bone marrow CFU were enriched up to 47% in mice infused with as

few as 5 x 104 transduced cells and selected with three rounds of BG and BCNU

treatments. Enrichment to 97% was obtained when 1 x 105 cells were infused prior to drug treatments. In vivo enrichment for MGMT-transduced human CD34+ cord blood

progenitors has also been achieved in NOD/SCID recipients preconditioned with

irradiation (Pollok et al. 2003) or a mild dose of BG and BCNU (Zielske et al. 2003).

The potency of MGMT-mediated stem cell selection has recently been demonstrated in a

large animal canine model (Neff et al. 2003; Neff et al. 2005). Transgene-expressing granulocytes were enriched to over 98% in both animals studied (from 3% and 16% initial cell expression percentages), following incremental dosing with BG and temozolomide. Remarkably, polyclonal marking and long-term expression was achieved with an average of only one integration event per cell.

The use of MGMT point mutants to differentially protect the hematopoietic compartment while sensitizing tumors has also been reported in animal xenograft models

(Koc et al. 1999; Reese et al. 1999; Kreklau et al. 2003). Clinical trials using gene transfer of MGMT have been proposed by several investigators, and one Phase I trial in patients with advanced malignancies such as melanoma, sarcoma and other solid tumors is in progress (Reese et al. 2004). The objective of this trial is to protect bone marrow stem cells from the toxic effects of chemotherapy and select for MGMT-G156A

21 transduced cells during treatment. This strategy is expected to result in less toxicity to bone marrow and blood cells while enriching for the number of genetically altered drug resistant stem cells over time, perhaps even from undetectable levels. The hypothesis of this study is based on preclinical data that shows this gene can provide HSCs with more than 500-fold survival advantage compared to HSCs not carrying the gene. In this clinical protocol, peripheral blood stem cells are collected from patients, exposed to a

Moloney murine leukemia retrovirus containing the G156A MGMT gene in the laboratory and immediately re-infused into the patient. Starting 2 days prior to cell infusion and every 6 weeks thereafter, patients are treated with BG and BCNU to inhibit tumor growth and provide selective resistance to the stem cells carrying the gene. To date, 5 patients have been enrolled and the level of gene transfer into the stem cells before infusion ranged from 11-36%. No complications related to cell infusion or chemotherapy administration have been observed. In one patient, evidence of genetically altered cells was observed by molecular analysis in the bone marrow 5 weeks after the infusion and prior to the chemotherapy treatment. While preliminary, these results indicate that infusion of hematopoietic stem cells transduced with retroviral mutant

MGMT is feasible and safe. This is an important trial because stem cell selection with

MGMT may be useful in other planned clinical applications including the use of MGMT in combination with therapeutic genes to correct for genetic disorders and the use of

MGMT stem cell protection during allogeneic transplantation as a selection strategy to encourage donor engraftment. Since the theoretical risks of oncogenesis associated with oncoretroviral vector integration has now been observed in a successful human gene therapy trial for SCID XI (Kaiser 2003), the next generation of gene therapy trials will

22 likely incorporate lentiviral vectors. These vectors are thought to have a decreased risk of insertional oncogenesis, and their increased stem cell transduction efficiencies indicate that lower multiplicities of infection (MOIs) may be used to achieve the same endpoint.

Nucleoside and Folate Analog Resistance

Several other drug resistance genes have been evaluated for in vivo selection or chemoprotection, including cytidine deaminase (CDA), and dihydrofolate reductase

(DHFR). CDA belongs to a class of enzymes involved in pyrimidine metabolism and salvage pathways. CDA can also inactivate cytosine nucleoside analogs, (e.g. cytarabine), which are used as antineoplastic agents. Murine bone marrow cells transduced with CDA was shown to increase hematopoietic CDA expression levels in recipient animals, but showed little evidence of drug protection (Eliopoulos et al. 1998).

Selection efficiency for CDA-transduced bone marrow cells ex vivo is highly dependent on cell density, suggesting that CDA released into the media may inactivate the drugs and protect neighboring cells (Beausejour et al. 2001).

DHFR converts folate into tetrahydrofolate, a cofactor required for thymidylate and purine biosynthesis. Folate analogs such as methotrexate (MTX) and trimetrexate

(TMTX) bind to DHFR with greater affinity, thereby inhibiting DNA synthesis (Fig. 5).

Specific DHFR mutations have been identified that are resistant to these antifolates (Zhao et al. 1994; Lewis et al. 1995). Early experiments showed that murine bone marrow transduced with DHFR mutants efficiently protected irradiated recipients from methotrexate-induced marrow toxicity (Corey et al. 1990; Zhao et al. 1994). Allay et al. demonstrated that nucleoside transport inhibitors, such as

23 nitrobenzylmercaptopurineriboside phosphate (NBMPR-P), increased the sensitivity of primitive hematopoietic cells to folate analogs (Allay et al. 1997).

Figure 5: The role of DHFR in pyrimidine biosynthesis. Antifolates (e.g. MTX) competitively bind to DHFR, blocking tetrahydrofolate synthesis. DHFR mutants (DHFR’) are resistant to antifolate inhibition. Nucleoside transporters allow cells to salvage nucleosides. NBMPR-P inhibits nucleoside transporters, further sensitizing untransduced cells to antifolate drugs.

Subsequent murine transplant experiments, using both TMTX and NBMPR-P for in vivo selection, resulted in a significant expansion of DHFR-transduced progenitors (Allay et al. 1998) However, only transient and limiting levels of DHFR-mediated enrichment was observed in nonhuman primate models in vivo, following combined TMTX and NBMPR-

P treatment (Persons et al. 2004). Increased toxicity was also evident in this model.

24 Further, pretreatment with cytokines failed to significantly increase the selection

stringency for long-term repopulating cells transduced with DHFR.

DUAL-GENE TRANSFER STRATEGIES

Although a major focus for in vivo HSC enrichment methods has been to reduce

myelosuppression, another emphasis has been to use this strategy to repopulate a diseased

hematopoietic compartment with gene-corrected cells. This application requires vectors

that are capable of efficiently expressing both the selectable marker and the therapeutic

gene. The selectable marker gene allows the transduced cells to be enriched, and the

therapeutic gene restores function to the diseased cells. Several vector designs have been generated for this endpoint.

Bicistronic vectors are constructed using internal ribosome entry site (IRES) sequences. If two genes are positioned in the vector in tandem they will be transcribed as a bicistronic message. However, in this configuration the first gene will be translated by

the normal cap-dependent mechanism, but the second gene will only extend the 3'

untranslated region. An IRES element positioned between the two genes will reinitiate translation to generate the second gene product. IRES elements were initially discovered in viral genomes, but have now been identified in many other organisms. The majority of the IRES sequences used for dual-gene vectors are derived from the polio virus or encephalomyocarditis virus genomes. IRES-initiated translation is typically less efficient than cap-mediated translation (Zhou et al. 1998; Yu et al. 2003). Thus, the orientation of the two genes in a bicistronic vector must be taken into account for each application. For instance, selection of cells transduced with a therapeutic-IRES-drug-resistance gene

25 configuration might lead to an oligoclonal population with high therapeutic gene expression, since high levels of the bicistronic transcript will compensate for low IRES translation rates. The same genes configured in the opposite orientation (with respect to the IRES) might result in a polyclonal population with low therapeutic gene expression.

IRES elements with tissue-specific or inducible expression patterns have also been identified, some of which are being developed for gene therapy applications (Creancier et al. 2000; Warnakulasuriyarachchi et al. 2004). Several gene therapy models have used

IRES elements to couple drug resistance genes to therapeutic or reporter gene expression.

Bicistronic vectors containing two separate drug resistance genes have also been used to expand HSC resistance to additional classes of chemotherapy drugs (Jelinek et al. 1999).

The foot and mouth disease virus (FMDV) 2a element has also been utilized in many dual-gene vectors (de Felipe et al. 1999; Milsom et al. 2004). The 54 bp sequence encoding FMDV-2a is positioned between two genes, the first of which has the stop codon removed. Thus, both genes and the 2a element are joined as one open reading frame. After the first gene and 2a sequence are translated, cis-acting hydrolase activity within the 2a residues cause ribosomes to "skip" the last peptide bond in 2a. Thus, the first gene product is released with 17 residues from the 2a element fused to its C- terminus. The ribosome then continues translating the second gene product, which contains an N-terminal proline from the 2a sequence (de Felipe 2004). Like IRES elements, the efficiency of ribosome slippage sequences appears to be sensitive to the specific gene combinations used (Milsom et al. 2004). Further, the activity of the first gene product can be perturbed by the 2a residues that remain fused to its C-terminus

(Lengler et al. 2005).

26 Alternative splicing mechanisms have also been utilized to create dual gene

vectors (Bowtell et al. 1988; Zhu et al. 2001). These vectors are constructed by adding

an extra splice acceptor site in front of the second gene. If the first splice acceptor is

recognized, the first gene is translated and the second gene becomes an extended 3'

untranslated region. If the second splice acceptor is recognized the first gene becomes

part of the excised intron and the second gene is translated. However, stringent selection

can result in the amplification of cells that preferentially splice out the therapeutic gene in

favor of drug resistance gene expression. Additional transcription cassettes can also be

inserted into vectors to co-express two genes. This strategy can be complicated by

transcriptional interference, resulting in the silencing of one gene in favor of the other

(Emerman et al. 1984). Recently Amendola et al. created lentivectors containing synthetic promoters that are bidirectional. These synthetic promoters allow two genes in opposite orientations to be expressed from the same promoter (Amendola et al. 2005).

Another strategy utilizes a mixture of separate single-gene lentiviral vectors to cotransduce cells (Fig. 6). This strategy allows the rapid evaluation of two genes without a need for dual-gene vector construction. Cotransduction with lentiviral vectors has only been described in vitro.(Frimpong et al. 2000; Reiser et al. 2000) In these studies separate VSVG-pseudotyped lentiviruses were shown to cotransduce cell lines or primary human neurons at a frequency proportional to the transduction efficiency of each virus.

Frimpong et al also reported efficient cotransduction of cells with two bicistronic vectors, each with a unique drug resistance gene, for in vitro selection of only dual positive cells.(Frimpong et al. 2000) Cotransduction of hematopoietic cells has not been evaluated. In the context of ex vivo gene therapy, the vast excess of unmodified

27 endogenous cells in vivo would limit the efficacy of selection prior to transplant. The

prolonged ex vivo culturing periods required for pre-selection have also been shown to

correlate with reduced cell pluripotency and engraftment.(Kittler et al. 1997; Gothot et al.

1998) Therefore, brief transduction schemes followed by in vivo enrichment has been a focus of most hematopoietic drug resistance gene transfer strategies.

Figure 6: Strategy for lentiviral cotransduction and selective enrichment. Cotransduction of cells with separate lentiviral vectors should generate populations that are untransduced, singly-transduced, or cotransduced. MGMT singly-transduced and cotransduced populations should be enriched equally with drug treatment, while untransduced and GFP singly-transduced cells are lost.

28 TARGETED TRANSDUCTION

Since there is an apparent tradeoff between specificity and transduction efficiency with natural viral proteins, many attempts to engineer viruses with specificity towards clinically relevant target cells have been made. Previous strategies have focused on localizing the MuLV-based vectors to target cells through the use of antibody bridges

(Roux et al. 1989; Etienne-Julan et al. 1992; Etienne-Julan et al. 1992) or envelopes that display ligands such as erythropoietin, heregulin, and hepatocyte growth factor (Kasahara

1994; Han et al. 1995; Nguyen et al. 1998). However these experiments required co- expression of the wild-type envelope, or only moderately improved the transduction efficiency.

Other strategies have utilized the insertion of sequences within the surface domain

(SU) of viral envelopes that force transduction through new receptors. These insertions are usually ligands (Cosset 1995; Nguyen et al. 1998) or single-chain antibodies with binding specificity for specific proteins expressed in the target cell (Russell 1993; Chu et al. 1995; Somia et al. 1995; Ager et al. 1996; Marin 1996; Jiang et al. 1998; Konishi

1998). Although many insertions have been shown to fold correctly and are presented on the virion surface, the transduction efficiency is often limited by perturbations in the transmembrane domain's (TM) fusion activity. Flexible linkers inserted between the ligand-SU junction have only moderately improved transduction efficiencies (Valsesia-

Wittmann et al. 1996). The process of infection involves SU trimerization, receptor binding, and conformational changes that expose TM for fusion. Thus, the insert must be presented in a way that redirects binding without perturbing SU trimerization or fusion.

Recently, Chandrashekran et al. demonstrated that expression of a membrane-bound form

29 of stem cell factor in an ecotropic packaging cell line resulted in virions containing the ecotropic envelope and stem cell factor on their surface, which were able to specifically

transduce human cells expressing the SCF receptor (Chandrashekran et al. 2004). Thus,

the ecotropic envelope may be able to utilize the human receptor or induce receptor-

independent fusion upon gaining entry into endosomes. Regardless, this study

demonstrates that envelope modifications may be unnecessary.

Additional strategies have involved the use of retroviral display by randomizing

the receptor binding domains and using entry-based screens to identify targeted virions

(Bupp et al. 2002; Roth et al. 2002). However, the titers obtained with retroviral vectors

restrict the size of the libraries that can be screened. Since the structure for many

envelope proteins remains elusive, strategies utilizing inserted targeting motifs may benefit from the combined use of randomization strategies. Randomization of the residues flanking the targeting motif might aid in the identification of a display configuration that allows both targeted binding and functional envelope fusion activities.

Alternatively, other viral envelope proteins that are much more tolerant of insertions may permit more efficient targeted delivery vehicles.

BALANCING SAFETY WITH EFFICACY

Great strides have been made in vector development with much of the focus aimed at increasing viral titers and tropism. While this improves the range of applications for a given system, it has also led to the trend of maximizing expression without regard for the cumulative number of integrations. Theoretical concerns over insertional mutations arising from gene therapy vectors have now been realized in three

30 patients enrolled in X-linked severe combined immuno-deficiency (X-SCID) gene therapy trials (Hacein-Bey-Abina et al. 2003). These adverse events have encouraged investigators to reevaluate copy number and weigh the risks of insertional mutagenesis for gene therapy models used for preclinical relevance. Woods et al. have demonstrated that even early hematopoietic progenitors are susceptible to multiple lentiviral integration events when transduced with high MOIs (Woods et al. 2003). The finding that transcriptionally active regions of the genome are hotspots for integration further emphasizes the need for caution (Laufs et al. 2003). These issues point to the need for standards, beyond expression level, for evaluating the potency of a given vector and delivery system. An index based on expression level per insertion may be one such standard, but would require further standardization of the techniques used for copy number analysis. Kustikova et al. have pointed to the importance for such a standard by demonstrating the nonlinear relationship between insertion and expression levels

(Kustikova et al. 2003). They concluded that expression-based gene transfer efficiencies should be targeted to 30% or less to attain the highest expression level with the fewest insertions. Standardized cell lines and protocols for determining viral titer are also nonexistent, as each group tends to have a different method for defining MOI. The use of virus-loaded fibronectin plates and spinoculation protocols for transduction further complicates the use of MOI as an informative index. Ultimately, the number of insertions per cell and total cell dose required for therapeutic efficacy will have to be established for clinical trial risk assessments. The statistical risk of insertional mutagenesis and the need for elevated titers could be reduced by minimizing the number

31 of non-target cell transductions. For this reason, continued emphasis should be placed on

the development of vectors with integration site or target cell specificity.

Conclusion

In summary, many advances have been, and continue to be, made in hematopoietic gene transfer technology. However, multiple factors must be taken into account for each application. Cis- and trans-acting vector determinants, such as promoters, dual gene linkage elements, and envelope pseudotypes, often vary in efficiency based on the genes used, the cell source, the intended cell targets, and the methods used for transduction. The risks associated with each disease should always outweigh those associated with therapy itself. Studies aimed at defining the general risk of insertional mutagenesis and in vivo selection will provide essential insight, but these risks should also be evaluated for each application.

32 CHAPTER 2

VIRAL VECTORS FOR MURINE HSC GENE THERAPY MODELS

Summary

Combined treatment with BG and BCNU has been shown to selectively enrich

MGMT-P140K transduced murine BM progenitor cells in vivo. Therefore, we set out to

evaluate whether dual-gene vectors that couple MGMT-mediated selection to therapeutic

gene expression could be used to repopulate dysfunctional hematopoietic tissue with

gene-corrected cells. A model of X-linked chronic granulomatous disease (CGD) was

chosen for these studies since as few as 10% functional cells can protect chimeric patients

from the phenotypic infections associated with this disease. CGD results from mutations

in any of the four subunits comprising NADPH oxidase. Granulocytes and monocytes

utilize NADPH oxidase to generate the free radicals needed for oxidative degradation of

microbes. Several problems were encountered using bicistronic vectors for this

application. Thus, the focus of our research was shifted away from a specific disease model towards optimizing dual-gene transfer and expression in murine hematopoietic cells. This chapter describes the problems associated with dual-gene vectors for the correction of CGD and subsequent experiments aimed at identifying the best gene transfer vectors for murine hematopoietic cell models.

33 Hypothesis

X-CGD cells transduced with dual-gene MGMT and gp91phox vectors will have restored

NADPH oxidase activity, and the gene-corrected cells will be enriched with BG and

BCNU treatment.

Results

Dual-gene Vectors for Chronic Granulomatous Disease

As a first approach to coupling MGMT-mediated selection to reconstituted

NADPH oxidase activity, the MGMT gene was transcriptionally linked to the gene correcting the X-linked form of CGD (gp91phox) using the EMCV-IRES element. Two bicistronic cassettes, each with the two genes in opposite orientations around the IRES, were inserted into a modified Moloney murine leukemia virus vector (MFG) optimized for efficient expression in hematopoietic cells (Robbins et al. 1998); MFG-phox-IRES-

MGMT (MFG-PIM) and MFG-MGMT-IRES-phox (MFG-MIP) (Fig. 7).

Figure 7: Bicistronic Retroviral Vectors. The MGMT and gp91phox genes were transcriptionally linked in two MuLV vectors (MFG) using the EMCV IRES element. The genes were configured in opposite orientations around the IRES in each construct.

34 The human myelomonoblastic cell line PLB-985 X-CGD, containing a targeted

disruption in the gp91phox loci (Zhen et al. 1993), was transduced with MFG-PIM or

MFG-MIP to evaluate whether these vectors would reconstitute NADPH oxidase activity

and provide resistance to BG and BCNU treatment. Equivalent percentages of

transduced cells were obtained with each vector, based on flow cytometric detection of

MGMT expression (Fig. 8). However, cells transduced with MFG-MIP expressed higher

levels of MGMT than those transduced with MFG-PIM.

Figure 8: IRES driven MGMT expression levels are reduced. PLB X-CGD cells were tansduced with the PIM or MIP vectors at an MOI of 10. Two days after trnasduction the cells were immunolabeled for MGMT detection.

The transduced cultures were treated with 2 μM BG and 0-6.25 μM BCNU and plated in methylcellulose to allow outgrowth of drug resistant colony-forming units (CFU). As shown in figure 9, MFG-MIP transduced CFU tolerated higher BCNU doses. Thus, despite having equivalent expression percentages, the lower IRES-mediated MGMT expression levels resulted in reduced protection from drug treatment.

35 Figure 9: Lower MGMT expression results in sensitivity to drug treatment. PLB X-CGD cells were mock transduced, or transduced with the MIP or PIM vectors (MOI = 10). Two days after transduction, the cells were untreated or treated with 2 μM BG and 0-15 μM BCNU and plated in methylcellulose. CFUs were enumerated 12 days after plating to calculate drug treatment survival percentages.

Control and drug-treated cultures were then treated with 0.5% dimethylformamide to

induce granulocytic differentiation (Tucker et al. 1987). The differentiated cultures were untreated or treated with phorbol ester 12-tetradecanoylphorbol-13 acetate (PMA) to

activate NADPH oxidase activity (Fig. 10). NADPH oxidase activity was visualized by

the addition of nitroblue tetrazolium (NBT), which forms a dark purple precipitate in the

presence of superoxide radicals. After drug treatment, the percentage of NBT-positive

cells increased in cultures transduced with either vector, but the NBT staining was less

36 intense in cells transduced with the MFG-MIP vector (Fig. 11). MGMT and gp91phox protein levels were further assessed by Western analysis (Fig. 12). Detection of both

proteins increased with increased doses of BCNU, but IRES-mediated expression levels

of either protein were lower at each BCNU dose, compared to cap-mediated translation

levels. Thus, IRES-mediated translation efficiency limited drug resistance or therapeutic

gene expression levels in cells transduced with the MFG-PIM and MFG-MIP vectors,

respectively.

Experiments were then carried out in a murine model of X-linked CGD to

determine if MFG-PIM transduced X-CGD cells could be enriched and reconstitute

NADPH oxidase activity in lethally irradiated murine recipients. The transplanted X-

CGD animals were highly sensitive to BG and BCNU doses that were tolerated in

previous selection experiments using the parental C57 mouse strain (Data not shown).

Drug sensitivity was attributed to low gene transfer efficiencies and reduced IRES-

mediated expression of MGMT. Although a few animals survived the drug treatments,

no evidence of reconstituted NADPH oxidase activity was detected. It was subsequently

reported that expression of vector-mediated expression of gp91phox in MuLV packaging

cell lines reduced the virus titers produced (Bellantuono et al. 2000). Therefore, we

decided to evaluate the use of lentiviral vectors for gene transfer and expression in

murine hematopoietic cells.

37 Figure 10: A human cell line model of X-CGD. A human myelomonoblastic cell line containing targeted deletions of the gp91phox loci (PLB X-CGD) were used to evaluate vector-mediated reconstitution of NADPH oxidase activity. PLB X-CGD cells were transduced with the MIP or PIM virus and the transduced cultures were differentiated into granulocytes with DMF. PMA was added to mature granulocyte cultures to induce assembly of the NADPH oxidase subunits, thereby activating the respiratory burst. Superoxide radicals were visualized by the formation of purple Nitroblue tetrazolium (NBT) precipitates.

38 Figure 11: Vector-mediated reconstituted of NADPH oxidase activity. NBT stained parental PLB (A) and PLB X-CGD (B) cells without (left) or with (right) PMA induction of NADPH oxidase activity. PLB X-CGD cells transduced with the PIM (C) or MIP (D) vectors were untreated (UT) or treated (T) with 2 μM BG and 6.25 μM BCNU. NADPH oxidase activity was visualized with NBT.

39 Figure 12: Dual-gene expression levels after drug treatment. MGMT and gp91phox expression levels in untreated or drug treated PLB X-CGD cells transduced with the MIP or PIM vectors. Cell extracts were derived from cultures represented in Figs 8-11.

40 Lentivector Expression in Hematopoietic Cells

Since low levels of gene transfer and IRES-mediated expression were obtained

with the murine retrovirus vectors, we evaluated the use of lentiviral vectors for gene

transfer and expression in hematopoietic cells. Three self-inactivating lentivirus

constructs were obtained, each utilizing different internal promoter sequences to drive

GFP expression; the cytomegalovirus (CMV) immediate-early promoter (pCMV-GFP),

the elongation factor 1-alpha (EF1α) promoter (pWPT-GFP), or a myeloproliferative

sarcoma virus promoter (MND), with a deleted negative control region and a substituted

primer-binding site (pMND-GFP). As a first approach to comparing the expression levels derived from each vector, the human erythroleukemia cell line, K562, was transduced with each virus using an MOI of 10. After 72 hours, the GFP expression levels in transduced cells were evaluated by flow cytometry. Over 98% of the K562 cells

in each culture expressed GFP. However, the mean GFP fluorescence intensity in K562 cells transduced with the MND-GFP vector was over 3-fold higher than those transduced

with the other lentiviral vectors (Fig. 13A).

Similar experiments were carried out on murine bone marrow mononucleocytes

(BM-MNCs), however, lower MOIs (= 0.5) were used in these experiments to prevent

multiple insertions from skewing the expression levels. As shown in figure 13B, CMV <

EF1α < MND, in the order of increased BM-MNC GFP expression levels. Although the

MND promoter resulted in the highest lentivirus-derived expression levels, the MFI of

GFP in BM-MNCs transduced with the MFG-GFP retroviral vector was still 5-fold

higher (data not shown). These data indicated that, among the lentivirus vectors

41 evaluated, the MND promoter expressed the highest level of GFP in K562 cells and

murine BM-MNCs.

A K562 B MURINE BM-MNCs

CMV EF1α MND CMV EF1α MND MFG*

Figure 13: Promoter strength comparison in hematopoietic cells. Lentivirus GFP expression vectors, each with a different internal promoter, were used to transduce K562 cells (A) or BM-MNCs (B). Plots represent the mean GFP fluorescence obtained after transduction with the pHR-cmvGFP (proCMV), pWPT-GFP (proEF1α), or pMND-GFP (proMPSV) lentiviral vectors. *An MFG-based MuLV vector was included for comparison to lentivirus vector expression levels

IRES-mediated expression levels are further reduced using lentiviral vectors

Although lower expression levels were obtained in hematopoietic cells using

lentiviral vectors, gene transfer efficiencies were much higher, compared to those

obtained retroviral vectors. Therefore, two dual-gene MGMT and gp91phox cassettes were assembled in the pMND lentiviral vector; an EMCV-IRES bicistronic vector containing gp91phox-IRES-MGMT (pMND-PIM), and a vector containing an FMDV-2a

42 slippage site linking the two genes (pMND-PAM). IRES-mediated MGMT expression

levels derived from MND-PIM transduced K562 cells were much lower than previous

levels obtained with the retroviral PIM vector, presumably due to lower lentivirus

transcription rates (data not shown). The 2a element in the MND-PAM vector resulted in

higher MGMT expression levels than any of the IRES-MGMT expression vectors, but

the 2a residues remaining on gp91phox abolished NADPH oxidase activity (data not

shown). Thus, the higher lentiviral titers allowed more efficient transduction rates, but

the reduced lentiviral vector expression levels exacerbated the problems associated with

dual-gene vector constructs.

High transduction and expression levels were obtained with both gp91phox and

MGMT single-gene MND vectors. Thus, subsequent studies were initiated to evaluate the ability to cotransduce cells with separate single-gene vectors, using MGMT-mediated selection to enrich dual-expressing cells. These results of these studies are presented in the following two chapters.

Materials and Methods

Vectors. Bicistronic retrovirus vectors, MFG-PIM and MFG-MIP: The cDNAs encoding the human genes for gp91phox (provided by Mary Dinauer, Indiana School of

Medicine) or MGMT-P140K were separately cloned into the MFG retroviral vector

(provided by J. Barranger, University of Pittsburgh) to generate MFG-phox and MFG-

MGMT, respectively. The EMCV-IRES element was obtained from the pCITE-2a vector

(Clontech). The IRES sequence was deleted from pCITE-2a by partial Pvu II and Msc I restriction and the vector was religated. 5'-NotI and 3'-Eco RI sites were added onto the

43 IRES sequence by PCR and the restricted IRES fragment was added back to the IRES-

deleted CITE vector, generating an IRES vector with both 5' and 3' multiple cloning sites,

pSOB. The gp91phox and MGMT cDNAs were then separately cloned into unique Nco I

and Not I sites of pSOB, to position the start codons of each gene at the site reported to

maximize IRES-mediated translation (Davies et al. 1992). The resulting vectors were

designated pSOB-phox and pSOB-MGMT, respectively. The IRES-MGMT cassette was

liberated from pSOB-MGMT using Bam HI and Bgl II restriction and was inserted into a

unique Bam HI site in MFG-phox to create pMFG-PIM. The same restriction sites were

used to liberate the IRES-gp91phox cassette from pSOB-phox for insertion into MFG-

MGMT to create pMFG-MIP.

Lentivirus vectors: Construction of the pHR-cmvGFP vector was described previously

(Zielske et al. 2003). The pWPT-GFP vector was obtained from Didier Trono

(University of Geneva, Switzerland).

pMND-GFP: A self-inactivating lentiviral luciferase vector, pCSO-rre-cppt-MCU3-

LUC, containing the MND promoter/enhancer sequences (Myeloproliferative sarcoma

virus enhancer, Negative control region deleted, dl587rev primer-binding site

substituted), the rev response element (rre) and the cPPT/CTS was obtained from Donald

Kohn (Halene et al. 1999). Luciferase was removed from this vector by partial restriction

with Nco I and Eco RI and replaced with a multiple cloning site, to generate pMND-

LINK. The cDNA encoding enhanced GFP was inserted into an Nco I site (closest to the

MND promoter) and a unique Bam HI site in pMND-LINK, to generate pMFG-GFP. pMND-PIM and pMND-MIP: The cDNA encoding human gp91phox was inserted into the Nco I site (closest to the MND promoter) and a unique Bam HI site. This construct

44 was then restricted at the unique Bam HI and Eco RI sites for insertion of the woodchuck

hepatitis virus posttranscriptional regulatory element (wPRE, provided by Thomas Hope,

University of Illinois at Chicago) to generate the pMND-phox single-gene vector. The

IRES-MGMT cassette was liberated from pSOB-MGMT using Bam HI and Bgl II

restriction and was inserted into the unique Bam HI site of pMND-phox, to generate

pMND-PIM. MND-PAM: The FMDV sequence was generated from oligos using

overlap extension PCR (Horton et al. 1989). A gp91phox-2a PCR fragment was

generated using the oligos, phox-FOR; 5'-tgcaataacgccaccaatctgaag-3' and phox-2A

REV; 5'-cctgccaacttgagcaggtcaaagctcaaaagctgtttcaccggtgcgaagttttccttg-3'. A separate 2a-

MGMT fragment was generated using the oligos MGMT-FOR; 5'- caagttggcaggggacgtcgagtccaaccctgggcctatggacaaggattgtgaa-3' and MGMT-REV; 5'- ttcatgggccagaagccatt-3'. The resulting phox-2a and 2a-MGMT fragments were annealed, extended, and amplified using the phox-FOR and MGMT-REV oligos. The resulting

fragment was cloned into pMND-PIM using unique Bcl I and Bsg I sites, generating

pMND-PAM. All constructs were verified by sequencing.

Virus production and transductions:

Retrovirus production: The pMFG-PIM and MFG-PIM vectors were transiently

transfected into ecotropic and amphotropic Phoenix packaging cell lines (ATCC). The

resulting ecotropic supernatants were used to transduce the stable GP+AM12 packaging

cell line and the amphotropic supernatants were used to transduce the stable GP+E86

packaging cell line (both obtained from Arthur Bank, Columbia University). Retrovirus

45 harvested from the stable ecotropic and amphotropic lines were ping-ponged onto each other to amplify the titers.

Lentivirus production: Virus was generated as described by Zielske et al.(Zielske et al.

2003) Briefly, 293T cells were co-transfected with the packaging vector

(pCMVdeltaR8.91), the VSVG pseudotyping vector (pMD.G), and the pMND- transducing vectors at a mass ratio of 3:1:3, using Lipofectamine 2000 (Invitrogen) according to the manufacturer's instructions. Virus produced 24-48 hours after transfection was harvested in Dulbecco’s modified Eagle medium (DMEM; Cellgro) containing 10% heat-inactivated FBS (Cellgro) and 2 mM GlutaMAX (Invitrogen).

Virus-enriched media was filtered through 0.45 μm syringe filter units (Millipore) and stored at -80° C. Expression titers were determined on K562 cells using virus dilutions that resulted in less than 10% transduction. Titers ranged from 0.5-3 x 107 expression units/ml. K562 cells were transduced in Iscove’s medium (Cellgro) containing 10% heat- inactivated FBS, 2 mM GlutaMAX, and 8 μg/ml polybrene (Sigma, St. Louis, MO).

PLB X-CGD cells (obtained from Mary Dinauer, Indiana School of Medicine) were transduced in RPMI (Cellgro) containing 10% heat-inactivated FBS, 2 mM GlutaMAX, and 8 μg/ml polybrene (Sigma, St. Louis, MO). BM-MNCs were transduced in alpha-

MEM containing 20% heat-inactivated FBS, 2 mM GlutaMAX, 6 μg/ml polybrene, in the presence of 20 ng/ml murine IL-3, and 50 ng/ml of murine IL-6, and SCF (R&D Systems

Inc., Minneapolis, Minnesota). MOIs defined for mouse cell transductions were based on expression titers established in K562 cells.

46 NADPH oxidase activity:

NBT assay: PLB and PLB X-CGD cells were differentiated into granulocytes by the addition of 0.5% dimethylformamide (DMF) to the growth media. After 5 days of differentiation, the cells were harvested, washed in serum-free RPMI, and suspended in

RPMI saturated with nitroblue tetrazolium. NADPH oxidase activity was induced by the addition of 100 ng/ml phorbol 12-myristate 13-acetate (PMA). The superoxide radicals react with NBT to form purple formazan deposits that can be visualized by microscopy.

DHR123 assay: NADPH oxidase activity was measured in PB and BM MNC samples using the dihydrorhodamine 123 assay, as described (Vowells et al. 1995). However, fluorescence was measured in the FL1 channel, according to the manufacturer's instructions (Molecular Probes), and granulocyte-specific NADPH oxidase activity was detected by co-labeling cells with the Gr-1 antibody (BD Biosciences).

Drug Selection: BG was synthesized by Dr Robert Moschel at the Frederick Cancer

Research Institute (Frederick, MD). BCNU was obtained from the Drug Synthesis and

Chemistry Branch of the National Cancer Institute (NCI; Bethesda, MD). All in vitro drug treatment incubations were carried out in serum free growth media at 37° C. Cells were pretreated with BG for 1 hour, followed by a 2 hour treatment in BCNU. CFU survival assays were performed as previously described (Davis et al. 1997). For in vivo selection, BG was dissolved to 3 mg/ml in 40% polyethylene glycol (Union Carbide

Corp., Danbury, Connecticut) and 60% PBS (pH 8.0), and BCNU was dissolved in ethanol and diluted to 1 mg/ml in PBS. Mice were injected intraperitoneally with 30

47 mg/kg BG, followed by 10 mg/kg BCNU an hour later. Three rounds of selection were

carried out at 3 week intervals, beginning at 3 weeks.

Flow cytometry: All flow cytometry results were obtained with an LSR flow cytometer

(Becton Dickinson). GFP expression was measured in cells after a single wash in PBS.

For MGMT detection, cells were fixed in 2% paraformaldehyde for 30 minutes at 4° C,

permeablized in 1% Tween-20 for 30 minutes at 37° C, and blocked in 10% normal goat

serum for 15 minutes at room temperature. The cells were then immunolabeled with the

mouse anti-human MGMT monoclonal antibody MT3.1 (Kamiya Biomedical, Seattle,

WA) and an (APC)-conjugated goat anti-mouse secondary antibody (Caltag Laboratories,

Burlingame, CA). Red blood cells were lysed in murine PB and BM samples prior to

preparation for flow cytometry.

48 CHAPTER 3

COTRANSDUCTION WITH SEPARATE SINGLE-GENE

LENTIVIRAL VECTORS

Summary

The P140K point mutant of MGMT allows robust drug selection of stem cells in vivo. This selection scheme has been coupled to therapeutic genes using dual-gene vectors. However, limited expression levels obtained with dual-gene vectors often require compensatory increases in the MOIs used. We evaluated whether hematopoietic cells could be efficiently cotransduced using low MOIs of two separate single-gene lentivectors, one of which contains MGMT, for dual-gene expressing cell enrichment.

Cotransduction efficiencies were evaluated using a range of MGMT:GFP vector MOI ratios, total MOIs, and selection stringencies. Cotransduction was optimal in vitro when equal proportions of each virus were used. Conversely, low MGMT:GFP vector MOI ratios resulted in the highest proportion of dual positive cells after selection. This strategy was then evaluated on murine bone marrow cells for in vivo selection. Equal virus proportions at low total vector MOIs (≤ 20) resulted in variable levels of dual-gene expressing cell engraftment. However, MGMT enrichment correlated with an increase in the percentage of GFP-expressing cells. Lower MGMT:GFP vector MOI ratios resulted in more consistent expression of both genes after drug treatment. Similar results were obtained for cells cotransduced with a single-gene MGMT and a single-gene, erythroblast-specific GFP vector. These data indicate that cotransduction could be used to efficiently couple stem cell selection to lineage-specific therapeutic gene expression.

49 Since insertional mutagenesis has been demonstrated as a potential risk associated with

gene therapy, these studies involved significantly reduced vector MOIs than previously

reported for in vivo selection with dual-gene lentiviral vectors.

Hypothesis

Separate single-gene lentiviral vectors, including MGMT, can be used to efficiently cotransduce hematopoietic progenitors cell populations, from which dual-gene expressing cells can be enriched with BG and BCNU treatment. Further, cotransducing cells with a ubiquitous MGMT expression vector and an erythroid-specific GFP expression vector

will restrict GFP expression to erythroid the progeny of MGMT-enriched progenitor

cells.

Results

Single Gene Vectors Efficiently Cotransduce Human Hematopoietic Cell lines

Selective expansion of hematopoietic cells cotransduced with two separate single- gene lentivectors, one of which provides drug resistance, has not been evaluated. As a first approach to evaluating the efficiency of this strategy, cotransduction and selection were carried out in vitro using the human K562 erythroleukemia cell line. A self- inactivating lentiviral vector containing the internal MND promoter was obtained from D.

Kohn (Halene et al. 1999). The wPRE and a multiple cloning site were introduced for subsequent generation of vectors expressing MGMT-P140K or the green fluorescence protein (GFP); pMND-MGMT and pMND-GFP, respectively. Both vectors express at high levels in K562 cells, as determined by flow cytometry. Although the GFP

50 fluorescence intensity is reduced when the cells are immunolabeled for MGMT

expression, distinct MGMT, GFP, and dual-gene expressing populations can be detected in cotransduced K562 cell populations (Fig. 14).

Figure 14: Dual detection of MGMT and GFP by flow cytometry. Representative flow cytometry plots showing MGMT and GFP expression in K562 cells transduced with MGMT and/or GFP vectors at an MOI of 0.25.

The cotransduction efficiency of K562 cells was first evaluated with the total vector MOI held at 0.5, varying the individual MGMT and GFP vector proportions (Fig.

15A). The degree of cotransduction was approximately equal to the product of the total

MGMT and total GFP expressing cell percentages. K562 cells cotransduced with an equivalent MGMT and GFP vector MOI mix (0.25:0.25) resulted in the highest MGMT- expressing and GFP-expressing cell percentages. This vector MOI mixture also resulted

51 in the highest percentage of dual-gene expressing cells (12%), which exceeded the

expected value (9%), based on the product of the total MGMT (29%) and total GFP

(31%) expressing cell percentages. The percentage of dual-gene expressing cells was lowest (4%) when an MGMT:GFP vector MOI mix of 0.05:0.45 was used. However,

cells transduced with this vector MOI mixture had the highest percentage of dual-gene

expressing cells (49%) after treatment with 10 μM BG and 15 μM BCNU (Fig. 15B).

Drug treatment enriched the percentage of MGMT-expressing cells to 95% in each

culture, with the level of enrichment inversely proportional to the MOI of the MGMT vector used. The MGMT-only and MGMT-GFP dual-gene expressing populations were expanded to the same extent, indicating that the level of MGMT expressed in both populations was equally protective. The percentage of GFP-expressing cells did not change after treatment since the proportion of GFP-expressing cells in the total culture is equivalent to the proportion of GFP-expressing cells in the drug resistant population. The initially high cotransduction rates achieved using an equivalent MOI of each vector ( =

0.25) had comparatively lower percentages of dual-expressing cells after selection (29%) due to the large proportion of MGMT-singly transduced cells present in the culture prior

to drug treatment. Thus, in the setting of drug selection, the highest percentages of dual-

gene expressing cells were obtained with virus mixtures composed of low MGMT and

high GFP vector MOI ratios.

52 Figure 15: Cotransduction and selective expansion levels using a constant MOI. K562 cells were cotransduced with varying MGMT:GFP virus ratios with the total MOI held at 0.5. (A) Total MGMT, total GFP, and dual positive percentages in the absence of selection. (B) Treatment with 10 μM BG and 15 μM BCNU enriched both MGMT only and MGMT-GFP dual positive populations. Total MGMT and total GFP values represent the net single and dual expression percentages for each gene. The expected level of dual positive cells is based on the product of the total MGMT and total GFP expression percentages for each data point. Error bars represent ± SEM, n= 3.

53 Slightly higher percentages of dual-gene expressing K562 cells were obtained than expected, based on the individual MGMT and GFP cell expression percentages.

Similar results were reported in a previous in vitro cotransduction study (Frimpong et al.

2000). To evaluate whether this was the result of virus complementation or the inability to completely resolve single-gene from dual-gene expressing populations, K562 cells were cotransduced with the GFP vector at an MOI of 0.25 and increasing MOIs (0 to 10) of the MGMT vector. The percentage of GFP-expressing cells was then evaluated by flow cytometry in the absence of MGMT staining. As shown in figure 16, the percentage of GFP-expressing cells was independent of the MGMT-expressing cell percentages

(measured separately), indicating that the higher than expected dual-gene expression percentages were due to the inability to completely resolve the single- and dual-gene expressing populations by flow cytometry.

Figure 16: MGMT and GFP transduction efficiencies are independent. MGMT and GFP expression levels measured separately. Transduction with GFP is unaffected ± by increased MOIs of the MGMT vector. Error bars represent SEM, n = 3

54 We next compared the efficiency of cotransduction in K562 cells using different total MOIs. Cotransduction of K562 cells with vector MOIs greater than 1 resulted in a high percentage of dual-gene expressing cells in the absence of selection; demonstrating that human cell lines are easily transduced with lentiviral vectors (data not shown).

However, high MOIs are associated with an increased risk of insertional mutagenesis.

We therefore evaluated whether cotransduction with an optimal MGMT:GFP (0.05:0.6) vector MOI ratio at a low total MOI would result in higher percentages of dual-gene expressing cells than cotransductions with higher total MOIs, but an equivalent

MGMT:GFP (0.5:0.5) vector MOI ratio. Before selection, the percentage of dual-gene expressing cells obtained using the higher total MOI was over 4 times that obtained with the 0.05:0.6 MGMT:GFP vector MOI ratio (21% versus 4.6% , respectively) (Fig. 17A).

However, after selection, cultures transduced with the 0.05:0.6 MGMT:GFP vector MOI ratio had the highest percentage of dual-gene expressing cells (66%), due to the increased selection pressure (Fig. 17B). These data indicate that selective expansion of cells cotransduced with optimal virus proportions can overcome the reduced transduction rates obtained with low total MOIs.

55 Figure 17: Cotransduction and selective expansion levels with varying MOIs. K562 cells were cotransduced with a low MGMT:GFP virus ratio or equal amounts of each virus at a higher total MOI. Single and dual positive MGMT and GFP expression levels were evaluated prior to (A) and after selection. with 10 μM BG and 15 μM BCNU (B). Total MGMT and total GFP values are the sum of the corresponding single and dual expressing populations. Fold enrichment of dual positive populations after drug selection are indicated for each transduction. Error bars indicate ± SEM, n = 3.

Drug Selection Enriches MGMT-Only and Dual-Positive Populations Equally

Drug selection of cells transduced with dual-gene vectors can lead to preferential expression of only the drug resistance marker (Hildinger et al. 1998). To evaluate whether cells that only express MGMT out-compete dual-gene expressing populations under stringent expansion conditions, the untreated cotransduction cultures from figure

17A were diluted into a 50 fold excess of untransduced K562 cells and selected with two sequential rounds of 10 μM BG and 15 μM BCNU treatment. As shown in figure 18, the

fraction of dual-gene expressing cells in the total MGMT-expressing population is

equivalent before and after two rounds of drug selection, demonstrating that preferential expansion of MGMT singly-expressing cells did not occur.

56 Figure 18: MGMT and dual-gene expressing cells are enriched equally. Cotransduced K562 cultures from Figure 17A were undiluted or diluted into a 50 fold excess of untransduced K562 cells and treated twice with 10 μM BG and 15 μM BCNU. The fraction of dual positive cells in the total MGMT-expressing population remains constant after stringent selection conditions. Error bars indicate ± SEM, n = 3.

In Vitro Selection Enriches Cotransduced Bone Marrow Progenitor Cells

As a first approach to evaluating cotransduction efficiency of primary cells, 5-

Fluorouracil enriched BM-MNCs were obtained. Higher MOIs were required to overcome the reduced transduction efficiency of primary murine cell populations, compared to human cell lines. BM-MNC cultures were cotransduced with a total vector

MOI of 20, using either equivalent (10:10) or staggered (5:15) MGMT:GFP vector MOI ratios. The transduced cultures were then treated with 20 μM BG and 0-25 μM BCNU, and expanded in culture for 10 days to determine the percentage of cells expressing each vector, or plated in methylcellulose to calculate the percentage of drug resistant hematopoietic progenitors. Prior to drug treatment, the dual-gene expressing cell

57 percentages were higher in BM-MNC cultures cotransduced with the equivalent

MGMT:GFP vector MOI ratios (27%) compared to cultures cotransduced with the staggered vector MOI ratio (20%) (Fig. 19A). However, 10 days after selection with 20

μM BG and 25 μM BCNU, the percentages of dual-gene expressing cells were highest in cultures cotransduced with the staggered (5:15) MGMT:GFP vector MOI ratios (67%) due to the greater selection pressure. The percentage of CFU in each culture that survived drug treatment was proportional to the MGMT vector MOIs used (Fig. 19B).

Transduction efficiencies are reduced in more primitive hematopoietic cell populations. To determine the cotransduction efficiency in more primitive populations, similar comparisons were carried out on Sca+/Kit+/Lineageneg (SKL) cells. SKL cell cultures were cotransduced using a total vector MOI of 100, but the individual vector

MOIs were similarly divided into equal (50:50) or staggered (20:80) MGMT:GFP vector

MOI ratios. The initial percentage of GFP-expressing cells in the cotransduced BM-

MNC cultures had a greater impact on the dual-gene expression percentages after drug treatment than the MGMT expression percentages, since even limiting numbers of drug resistance cells could be efficiently enriched. Therefore, the staggered MGMT:GFP vector MOI ratio was further skewed in these experiments, in favor of GFP. Higher

MOIs were required to efficiently transduce SKL cells, compared to BM-MNCs. As an indicator of the overall lower transduction rate of SKL cells, the 5-fold higher dual-gene vector MOIs used to transduce SKL cells resulted in dual-gene expressing cell percentages that were similar to those obtained in BM-MNCs transduced with lower

MOIs (Fig. 20A compared to Fig. 19A).

58 A.

B.

Figure 19: Selective expansion of cotranduced murine bone marrow cells. 5-FU enriched whole BM cells were cotransduced for 12 hours using equivalent (10-10) or staggered (5-15) MGMT:GFP virus mixtures. (A) Single and dual-positive expression percentages in culture expanded bone marrow before or after drug treatment with 20 μM BG and 25 μM BCNU. (B). The percentage of CFU surviving treatment with 20 μM BG and 0-30 μM BCNU. Error bars indicate ± SEM, n = 3.

59 Again, the highest post-treatment percentage of dual-gene expressing cells (73%) was

observed cultures cotransduced with the staggered (20:80) MGMT:GFP vector MOI

ration, due to the greater selection pressure.

MGMT expression predicted CFU drug resistance in the SKL cell cultures but not in the BM-MNC cultures, presumably due to the variability of transduction rates among

different cell types. Thus, in transduced SKL cell cultures, the initial percentage of

MGMT expression correlated with CFU survival after drug exposure (Fig. 20B), whereas

CFU survival in the drug treated BM-MNC populations were much lower than the

MGMT expression percentages in the liquid cultures (Fig. 19B).

60 A.

B.

Figure 20: Selective expansion of cotranduced murine SKL cells. SKL cells isolated from whole BM were cotransduced for 12 hours using equivalent (50-50) or staggered (20-80) MGMT:GFP virus mixtures. (A) Single and dual-positive expression percentages in culture expanded populations before or after drug treatment with 20 μM BG and 25 μM BCNU. (B). The percentage of CFU surviving treatment with 20 μM BG and 0-30 μM BCNU. Error bars indicate ± SEM, n = 3.

61 Cotransduced Hematopoietic Progenitor Cells are Enriched In Vivo with Drug

Selection

Although transduction of quiescent HSCs is more efficient with lentiviral vectors,

high MOIs are routinely used for gene transfer studies in murine BM cells. Our in vitro cotransduction results demonstrated that the reduced transduction efficiency achieved with lower MOIs of the MGMT virus can be counterbalanced by increased selection pressure. However, the in vitro results are only indicative of the progenitor population.

The transduction efficiency of the pluripotent stem cell population is expected to be further reduced, compared to that of the more mature hematopoietic progenitors.

Therefore, a high degree of variability was expected for vector-expressing cell engraftment and expansion in vivo.

As a first approach to evaluating the cotransduction strategy for in vivo selection,

BM-MNCs from 5-FU treated mice were cotransduced for 12 hours and transplanted into

6 lethally-irradiated recipients. Equivalent MGMT and GFP vector MOIs (=10, each) were used to cotransduce the donor marrow to increase the initial number of cotransduced progenitors available for selective expansion. The percentages of MGMT

(31%) and GFP (29%) expressing cells were measured separately in a fraction of culture- expanded donor cells (data not shown). After a 3 week engraftment period, the mice were treated with or without regimens of 30 mg/kg BG and 10 mg/kg BCNU. Drug selected cohorts received 3 treatments separated by 3 week intervals. After recovery, and

2 days prior to subsequent treatment, MGMT- and GFP-expressing peripheral blood mononucleocyte (PB-MNCs) percentages were measured by flow cytometry. The mice were sacrificed at 15 weeks to assess the percentage of BM-MNCs expressing each gene.

62 MGMT GFP MGMT GFP MGMT GFP MGMT GFP MGMT GFP MGMT GFP

Figure 21: Cotransduction and in vivo selection of murine BM-MNCs. 5-FU enriched BM cells were cotransduced with MND-MGMT and MND-GFP vectors (MOI = 10, each) prior to transplant into lethally irradiated recipients. PB MGMT and GFP expression precentages were taken at three week intervals, two days prior to each treatment with 30 mg/kg BG and 10 mg/kg BCNU. The plot represents MGMT+ and GFP+ expressing PB-MNC percentages at weeks 3 and 12, and BM-MNC expression percentages at sacrifice (week 15). Individual mice were untreated or treated three times with BG and BCNU.

Three weeks after transplant, the range of MGMT-expressing PB-MNCs in the cohort of 6 mice was 17-32%, while GFP-expressing cells ranged from 9-28% (Fig. 21).

MGMT and GFP expression percentages declined without selection, with only 2 of 3 untreated control animals showing evidence of expression in the BM-MNCs at the time of sacrifice (10-18% MGMT, no GFP). Of 3 drug treated animals, 2 engrafted with vector-expressing BM-MNCs after 15 weeks; 1 animal had a low percentage of MGMT- expressing (2%) and GFP-expressing (1%) cells, while the other animal had high BM-

MNC expression percentages of both MGMT (51%) and GFP (94%) cells. Although

63 MGMT expression was evident in the third animal's PB-MNCs for up to 12 weeks, no

MGMT-expression was detected in the BM-MNCs. BM-MNCs from the animal with the highest MGMT and GFP expression percentage revealed that 75% of lymphoid cells

(B220+ and CD3+) and 99% of the myeloid cells (gran-1+) expressed GFP; indicating

enrichment of cotransduced progenitors was achieved (data not shown). MGMT and

GFP copy number analysis on CFU derived from this animal revealed an average of 1

copy of each vector per cell (data not shown). Thus, the degree of dual-positive cell

engraftment and enrichment among animals is variable when the cotransduced stem cell

pool is limiting.

The in vitro cotransduction studies demonstrated that the initially low dual-gene

expression percentages obtained with low MGMT:GFP vector MOI ratios resulted in the

highest percentage of dual-expressing cells after selection. Therefore, this strategy was

applied to a second transplant experiment. The same total MOI (= 20) was used in this

experiment, but the MGMT:GFP vector MOI mixture was changed to 5:15. SKL cells

were used to decrease variability among recipients and increase the percentage of

cotransduced progenitor cells available for selective expansion. Donor cells were

transduced for 12 hours prior to transplant into 8 lethally irradiated recipients. Each

recipient animal was transplanted with 1,000 of the virus-exposed SKL cells and 2 x 106 lineage-positive BM-MNCs for hematopoietic support. Although a higher apparent GFP vector MOI was used, the total MGMT- and GFP expression percentages in culture expanded cells was 80% and 32%, respectively (data not shown).

64 MGMT GFP MGMT GFP MGMT GFP MGMT GFP MGMT GFP MGMT GFP MGMT GFP MGMT GFP

Figure 22: Cotransduction and in vivo selection of murine SKL cells. SKL cells isolated from whole BM were cotransduced with MND-MGMT and MND-GFP vectors (MOI = 5 and 15, respectively) for 12 hours prior to transplant into lethally irradiated recipients. PB MGMT and GFP expression precentages were taken at three week intervals, two days prior to each treatment with 30 mg/kg BG and 10 mg/kg BCNU. The plot represents MGMT+ and GFP+ expressing PB-MNC percntages at weeks 5 and 14, and BM-MNC expression percentages at sacrifice (week 17). Individual mice were untreated or treated three times with BG and BCNU.

Five weeks after transplant, the percentage of MGMT-expressing PB-MNCs in

the cohort of 8 untreated animals ranged from 7-42% and the percentage of GFP-

expressing cells ranged from 2-37% (Fig. 22). Six of the 8 mice were then drug treated,

as described above. In the 2 untreated controls; one animal had low MGMT and high

GFP-expressing PB-MNC percentages after 14 weeks, while the other control animal had

a steady increase in GFP-expressing PB-MNCs and a total loss of MGMT detection over

the same time period. Although drug treatments correlated with increased MGMT-

expression (P = 0.004) and GFP-expression (P = 0.043) percentages in PB-MNCs, a wide

65 range of MGMT (38-73%) and GFP (2-53%) expression percentages were detected in the

PB-MNCs among the 6 animals subjected to 3 rounds of selection. Of the 6 drug treated

animals, 2 had high percentages of MGMT and low percentages of GFP in their BM-

MNCs, 3 animals had moderate BM-MNC percentages of each gene, and one animal had both high MGMT and GFP BM-MNC expression percentages at the time of sacrifice.

Real-time PCR analysis on CFU from these animals revealed copy numbers clustering

around 0, 1, or 2 copies of each vector per cell (Fig. 23). However, the animal (#8) with

the highest MGMT-expressing cell percentages also had a higher average MGMT copy

number per cell (= 8 ± 2). These data indicate that the use of stem cell-enriched donor

populations increased engraftment and selective expansion of MGMT-transduced cells in

vivo without a need for high levels of vector insertions. In addition, selective expansion for MGMT-expressing cells corresponded with increased GFP-expressing BM-MNC percentages, despite the lower transduction efficiency obtained with the GFP virus.

66 Figure 23: Vector insertions per cell after in vivo selection. Analysis of MGMT and GFP vector copy numbers per diploid genome (cell) in BM CFU derived from recipient animals in Figure 22.

67 Cotransduction for In Vivo Selection and Lineage Specific Reporter Gene Expression

Vector size constraints can be a limiting factor for dual-gene vectors. Insertion of

two large genes into the same vector can reduce titers and leave little space for other beneficial sequences, such as post-transcriptional regulatory elements, insulators, or matrix attachment regions (Kumar et al. 2001). Cotransduction may be particularly suitable for situations in which the transcription of two genes is targeted to different cell types. To evaluate whether cotransduction with a ubiquitous MGMT expression vector and a lineage-restricted GFP virus would allow both stem cell selection and expression of

GFP in the correct hematopoietic compartment, we used a modified lentiviral GFP reporter vector (pRRL-GATA-GFP) that restricts GFP expression to TER119+ murine

erythroblasts (Lotti et al. 2002). SKL cells were cotransduced with an MGMT:GATA-

GFP vector MOI mixture of 5:15, and transplanted into 6 lethally-irradiated recipients, as

described above.

At five weeks post-transplant MGMT-expressing PB-MNC percentages in the

cohort of 6 untreated animals averaged 19% (6-51%), while GFP-expression percentages

in TER119+ erythroblasts averaged 18% (9-29%) (Fig. 24). MGMT-expressing PB-MNC

percentages increased to an average of 41% (range 14-65%) in the 4 animals selected

with three rounds of BG and BCNU treatment, compared to 2-7% in 2 untreated controls.

Contaminating erythrocytes in the PB-MNC samples elevated Ter119+ counts thereby

reducing the apparent GFP-expressing erythroblast percentages. Nevertheless, the

percentage of GFP-expressing cells in the TER119+ BM-MNC fraction averaged 57%

(range 1-96%) in the 4 drug selected animals, compared to 3-44% in the 2 untreated

controls. The two animals with the highest MGMT-expressing cell enrichment also had

68 MGMT GFP MGMT GFP MGMT GFP MGMT GFP MGMT GFP MGMT GFP

Figure 24: Cotransduction links selection and lineage specific expression in vivo. Murine SKL cells were cotransduced with MGMT and RRL-GATA-GFP viruses (MOI = 5 and 15, respectively) for 12 hours prior to transplant into lethally irradiated recipients. PB MGMT+ and TER119+-GFP+ expression percentages were taken at three week intervals, two days prior to each treatment with 30 mg/kg BG and 10 mg/kg BCNU. The plot represents MGMT+ and GFP+ expressing PB-MNC percentages at weeks 5 and 14, and BM-MNC expression percentages at sacrifice (week 17). Individual mice were untreated or treated three times with BG and BCNU.

the highest percentage of GFP-expression in the TER119+ BM-MNC fraction (Fig. 25).

As expected, GFP expression was restricted to TER119+_erythroblasts (Fig. 26). Copy

number analysis revealed an average of 1-4 copies of each vector per cell (Fig. 27). Thus,

cotransduction allows lineage-specific gene expression to be coupled to progenitor

selection in vivo.

69

Figure 25: GFP expression in recipient TER119+-BM-MNCs. Flow cytometric analysis of TER119+-specific GFP expression percentages in BM-MNCs dervived from untreated and drug treated recipient animals depicted in Figure 24.

70 Figure 26: Gata-GFP expression is limited to Ter119+ erythroblasts. BM from animal #6 in Figure 24 was stained with the Ter119+ (erythroblasts), Gr-1 (granulocytes), or B220 (B-cell) antibodies for lineage-specific evaluation of GFP expression.

71 Figure 27: Vector copy numbers per cell after in vivo selection. Analysis of MGMT and RRL-GATA-GFP vector copy numbers per diploid genome (cell) in BM CFU derived from recipient animals in Figure 24.

A second transplant experiment was carried out to determine whether this strategy is efficient with limiting MND-MGMT and GATA-GFP vector MOIs. BM-MNCs harvested from 5-FU treated mice were transduced with each vector at an MOI of 3.

Following a 12 hour transduction, 2 x106 cells were transplanted into 6 lethally-irradiated

recipients. At three weeks post-transplant, the percentage of MGMT-expressing PB-

MNCs in the cohort of 6 untreated animals ranged from 6-27%, while little to no GFP expression was detected (Fig. 28). Of the four animals selected with three rounds of 30 mg/kg BG and 10 mg/kg BCNU selection, only two showed evidence of gene-expressing cell engraftment. The percentage of MGMT-expressing PB-MNCs in these two animals was 56-72% after selection, compared to 6-13% in the untreated controls. In addition,

72 the percentage of GFP-expressing erythroblasts was 26-42% in the drug treated animals, compared to 3-5% in the untreated animals. The mice were sacrificed at 15 weeks post- transplant. The two drug treated animals had high percentages of both MGMT- expressing BM-MNCs (55% and 69%) and GFP-expressing erythroblasts (56% and

74%). In CFU derived from these animals, MGMT insertions ranged from 1-3 copies per cell and GFP vector insertions ranged from 1-4 copies per cell. These data indicate that

MGMT-mediated selection could easily be coupled to a range of lineage-specific therapeutic vectors using the cotransduction strategy.

73 MGMT GFP MGMT GFP MGMT GFP MGMT GFP

Figure 28: Enrichment and lineage-specific expression in vivo with limiting MOIs. 5-FU enriched murine BM cells were cotransduced with MGMT and RRL-GATA-GFP viruses (MOI = 3, each) prior to transplant into lethally irradiated recipients. PB MGMT+ and TER119+-GFP+ expression percentages were taken at three week intervals, two days prior to each treatment with 30 mg/kg BG and 10 mg/kg BCNU. The plot represents MGMT+ and TER119+-GFP+ expressing PB-MNC percentages at weeks 5 and 14, and BM-MNC expression percentages at sacrifice (week 17). Individual mice were untreated or treated three times with BG and BCNU.

74 DISCUSSION

We evaluated whether cotransduction of hematopoietic cells with two separate

single-gene vectors, one of which allows selection, would provide an efficient alternative

to dual-gene vectors. These studies involved significantly reduced MOIs than previously

reported for in vivo selection with dual-gene bicistronic vectors. We found that: 1) the

percentage of dual-gene expressing cells following cotransduction is proportional to the

transduction efficiencies of each vector; 2) at limiting total MOIs, low MGMT:GFP

vector MOI ratios resulted in the highest level of dual-gene expressing cells after drug

selection; 3) cotransduction with single-gene vectors allows MGMT-mediated

progenitor cell expansion to be coupled to lineage-specific transgene expression. The use

of single-gene vector cotransduction for dual-gene delivery eases vector insertion size limits, and cis-acting complications that limit the expression efficiency of one or both genes in dual gene constructs.

Given that cotransduction involves two genes delivered in separate vectors, we evaluated different vector MOI ratios, and total MOIs, to optimize the percentage of dual- expressing cells in the setting of drug selection. When the total vector MOI was low and held constant, the percentage of dual-gene expressing K562 cells was found to be proportional to the MGMT and GFP expression percentages. The product of fractional proportions is highest when the proportions are equal. Therefore, equivalent levels of each vector resulted in the highest percentage of dual-gene expressing cells in the absence of selection. With drug treatment the initial percentage of GFP-expressing cells was most relevant, since the percentage of GFP-expressing cells in the total population is equivalent to the proportion of GFP-expressing cells in the drug resistant (MGMT-

75 expressing) population. Therefore, at a constant total MOI, low MGMT:GFP vector MOI

ratios resulted in the highest percentage of dual-gene expressing cells after selection. As

previously reported with other vectors (Frimpong et al. 2000), the percentage of dual-

gene expressing cells following cotransduction was slightly higher than the expected

values based on the individual gene expression percentages. However, in the absence of

MGMT staining, the percentage of GFP-expressing cells unaffected by increasing MOIs

of the MGMT virus. Thus the slightly higher dual-gene expression percentages observed

when the percentages of MGMT and GFP expression were measured simultaneously are

likely due to the inability to completely resolve the single- and dual-gene expressing

populations by flow cytometry.

The selection pressure must be taken into account when enriching cells that are cotransduced with low vector MOIs. The degree of MGMT-mediated cell expansion obtained with higher selection pressure can overcome the limiting percentage MGMT- expressing cells obtained with low MGMT vector MOIs. Low MGMT:GFP vector MOI ratios at a low total MOI resulted in a higher percentage of dual-gene expressing cells after selection than that obtained with equivalent MGMT:GFP vector MOI ratios at a higher total MOI. For dual-gene vectors, highly stringent drug treatments have been shown to preferentially select for drug resistance gene expression at the expense of the second gene (Hildinger et al. 1998; Kane et al. 2001). However, using a two vector system, MGMT and GFP dual-gene expressing cells were enriched to the same extent as those that only expressed MGMT, even after stringent selection conditions.

The efficiency of cotransduction for dual-gene transfer into primary BM-MNCs is more difficult to predict since it consists of a mixed population of cell types. Higher total

76 vector MOIs were used for these studies than for human cell lines to overcome the

reduced transduction rates obtained with murine BM-MNCs. Although the initially low,

the percentage of dual-gene expressing BM-MNCs obtained after drug treatment was

over two times higher than expected, based on the individual MGMT and GFP expression percentages. These data indicate that specific cells may be more or less susceptible to transduction and thus, the high vector MOIs routinely used for efficient transduction of primary cells will likely result in higher integration rates in the susceptible cell populations.

Since the efficiency of cotransduction with single-gene vectors is dependent on the transduction efficiency of each vector, we expected a low percentage of dual-gene expressing stem cells in the absence of selection. Therefore, equivalent vector MOIs were used for donor cell populations containing fewer stem cells (5-FU enriched BM-

MNCs) to increase the total number of dual-gene expressing cells available for expansion. In contrast, low MGMT:GFP vector MOI ratios were used to cotransduce

SKL cells since this cell fraction is highly enriched for stem cells. Although recipient mice engrafted with variable MGMT and GFP expression percentages, expansion of

MGMT-expressing cells in drug treated animals correlated with increased percentages

GFP-expressing cells.

Expression levels varied among animals regardless of the donor cell populations used for the transplants. Whereas variability seen in animals transplanted with 5-FU enriched BM-MNCs was likely due to limiting numbers of transduced stem cells, variability among animals transplanted with transduced SKL cells was likely due to differences in the populations that engrafted in each animal. Fewer mice transplanted

77 with 5-FU enriched BM-MNC populations showed evidence of gene-expressing cell

engraftment, but robust MGMT-mediated enrichment correlated with increased

percentages GFP-expressing cells in the drug treated cohorts. Interestingly, more of the

animals transplanted with SKL cells engrafted with MGMT-expressing cells, but the degree of enrichment was not increased. The presence of large numbers of MGMT- expressing progenitors in the engrafted population may limit the selection stringency at the stem cell level.

Cotransduction may have the most potential for situations in which the two genes are transcriptionally targeted to different cell types. This is especially true for severe hemoglobinopathies, including beta-thalassemia major and sickle cell anemia, in which bulky regulatory elements are required for lineage restricted therapeutic gene expression

(Leboulch et al. 1994; Pawliuk et al. 2001). We report here the ability to couple MGMT- mediated progenitor cell selection to erythroblast-specific GFP expression using cotransduction with separate single-gene vectors. Selection for engrafted MGMT and

GATA-GFP expressing cells was efficient using both SKL cells and 5-FU enriched BM-

MNCs as donor populations.

Cotransduction with single-gene vectors is an efficient alternative to dual-gene vectors. This strategy should be useful for evaluating the therapeutic potential of low copy gene replacement used to correct hematopoietic disease models. Since the gene expression levels are independent with this strategy, a wide range of therapeutic vectors could be rapidly assessed using pre-established selection conditions for the single-gene

MGMT vector. Although cotransduction requires at least two insertions for dual-gene expression, the enhanced expression efficiency achieved with single-gene vectors

78 suggests that fewer total copies may be required to achieve expression levels equivalent

to that of dual-gene vectors.

MATERIALS AND METHODS

Vectors and virus production: The self-inactivating lentiviral vector (pCSO-rre-cppt-

MCU3-Luc) containing an internal MND promoter, and the central polypurine tract/central termination sequence (cPPT/CTS) was kindly provided by D. Kohn (Halene et al. 1999). The luciferase gene was removed from this vector by Nco I/Eco RI digestion and replaced with a multiple cloning site cassette. The woodchuck hepatitis

virus posttranscriptional regulatory element (obtained from T. Hope) was inserted into the unique Eco RI site. The resulting vector was restricted with Nco I/Bam HI for insertion of MGMT-P140K or GFP. The pRRL-GATA-GFP vector was described previously (Lotti et al. 2002). Virus was generated as described by Zielske et al. (Zielske et al. 2003). Briefly, 293T cells were co-transfected with the packaging vector

(pCMVdeltaR8.91), the VSVG pseudotyping vector (pMD.G), and either the pMND-

MGMT-P140Kpre, pMND-GFPpre or pRRL-GATA-GFP transfer vectors at a mass ratio of 3:1:3, using Lipofectamine 2000 (Invitrogen) according to the manufacturer's instructions. Virus produced 24-48 hours after transfection was harvested in Dulbecco’s modified Eagle medium (DMEM; Cellgro) containing 10% heat-inactivated FBS

(Cellgro) and 2 mM GlutaMAX (Invitrogen). Virus-enriched media was filtered through

0.45 μm syringe filter units (Millipore) and stored at -80° C. Expression titers were determined on K562 cells using virus dilutions that resulted in less than 10% transduction. Titers ranged from 6x106 to 3x107 expression units/ml.

79

Cotransduction and Transplants: Virus preps were premixed at the specified MOI ratios prior to the addition of cells. K562 cells were cotransduced in Iscove's Modified

Dulbecco's Medium (IMDM, Cellgro) containing 10% heat-inactivated FBS, 2 mM

GlutaMAX, and 8 μg/ml polybrene (Sigma, St. Louis, MO). Whole BM and sorted stem cell populations were transduced for 12 hrs in alpha-MEM containing 20% heat- inactivated FBS, 2 mM GlutaMAX, 6 μg/ml polybrene, in the presence of 20 ng/ml murine IL-3, and 50 ng/ml of murine IL-6, and SCF (R&D Systems Inc., Minneapolis,

Minnesota). The apparent MOIs used for mouse cell transductions were based on expression titers established in K562 cells. The transduced cells were transplanted immediately after transduction. For whole BM transplants, donor marrow, was obtained from 6- to 8-week-old C57Bl/6j mice (Jackson Laboratories, Bar Harbor, Maine) 2 days after treatment with 150 mg/Kg 5-FU. SKL cells were isolated as previously described

(Reese et al. 2003). Six-week-old recipient mice were lethally irradiated with 850 cGy using a Cs137 source and transplanted with 2 x106 5-FU enriched cells per mouse, or

1,000 SKL cells supported with 2 x106 lineage-positive cells per mouse.

Drug selection: BG was synthesized by Dr Robert Moschel at the Frederick Cancer

Research Institute (Frederick, MD), and BCNU was obtained from the Drug Synthesis and Chemistry Branch of the National Cancer Institute (NCI; Bethesda, MD). All in vitro drug treatment incubations were carried out in serum free media at 37° C. Cells were pretreated with BG for 1 hour, followed by a 2 hour treatment in BCNU. K562 cells were treated in IMDM. Drug treatments for in vitro selection of transduced murine BM-

80 MNCs were carried out with 25 μM BG and 0, 5, 15, or 30 μM BCNU in alpha-MEM

containing 1.2% spleen cell conditioned media (Stem Cell Technologies). Drug treated

murine BM-MNC samples were expanded 10 days in liquid culture using the same media

formulation as described for transduction, without polybrene. Progenitor survival assays

were performed as previously described.(Davis et al. 1997) For in vivo selection, BG was dissolved to 3 mg/ml in 40% polyethylene glycol (Union Carbide Corp., Danbury,

Connecticut) and 60% PBS (pH 8.0), and BCNU was dissolved in ethanol and diluted to

1 mg/ml in PBS. Mice were injected intraperitoneally with 30 mg/kg BG, followed by 10

mg/kg BCNU an hour later. Three rounds of selection were carried out at 3 week

intervals, beginning at 3 weeks for animals transplanted with whole BM, and 5 weeks for

animals transplanted with SKL cells.

Flow cytometry: All flow cytometry results were obtained with an LSR flow cytometer

(Becton Dickinson). GFP expression was measured in K562 cells after a single wash in

PBS. For AGT detection, K562 cells were fixed in 2% paraformaldehyde for 30 minutes

at 4° C, permeablized in 1% Tween-20 for 30 minutes at 37° C, and blocked in 10%

normal goat serum for 15 minutes at room temperature. The cells were then

immunolabeled with the mouse anti-human MGMT monoclonal antibody MT3.1

(Kamiya Biomedical, Seattle, WA) and an (APC)-conjugated goat anti-mouse secondary

antibody (Caltag Laboratories, Burlingame, CA). Dual detection of AGT and GFP

expression was carried out with the same staining protocol as that used for AGT alone.

Red blood cells were lysed in murine PB and BM MNC samples prior to preparation for

flow cytometry.

81 Copy number analysis: Vector insertions per cell were evaluated using a LightCycler instrument and the LightCycler FastStart DNA Master SYBR Green I kit from Roche

Diagnostics (Basel, Switzerland). Vector copy numbers were determined using proviral- specific MGMT primers; Sense 5'- ACGTCTATATCATGGCCG-3', Antisense 5'-

TGAGGATCTTGACAGGATT-3', and GFP primers; Sense 5'-

ACGTCTATATCATGGCCG-3', and Antisense 5'-TGTGATCGCGCTTCTC-3'.

Genomic copy numbers were determined using GAPDL4-specific primers, as previously described.(Zielske et al. 2003) MGMT, GFP, and GAPDL4 copy numbers were assessed using a standard curve consisting of murine genomic PB-MNC DNA spiked with known vector copy numbers, and verified using K562 cell clones containing a single MGMT-

GFP vector insertion. One diploid copy of the mouse genome was assumed to contain two copies of GAPDL4 and 5.9 pg of DNA. Ten CFU were analyzed per animal and the average copy number per CFU was compared to that of total BM-MNC fraction from each animal.

82 CHAPTER 4

COMPARATIVE ANALYSIS OF DUAL-GENE DELIVERY STRATEGIES

Summary

Dual-gene vectors that couple therapeutic gene expression to MGMT-mediated stem cell selection allow reconstitution of diseased hematopoietic tissue with gene- corrected cells. However, dual-gene vectors are often limited by size constraints or poor expression levels. We previously evaluated the efficiency of cotransducing cells with separate single-gene MGMT and GFP lentivectors, using MGMT-mediated resistance to

BG and BCNU treatment to selectively enrich dual-gene expressing cells. In the current study we compared the efficacy of the cotransduction strategy to that of dual-gene vector strategies, using drug resistance to enrich dual-gene expressing populations. For the latter, we generated MGMT-GFP vectors containing either the EMCV-IRES or the

FMDV-2a element for co-expression. Each strategy was evaluated in vitro and in vivo using equivalent MOIs to transduce 5-FU or Sca+/Kit+/Linneg-enriched murine bone marrow cells. The highest dual-gene expression percentages were obtained with the

FMDV-2a dual-gene vector, but half of the gene products were fusion proteins.

Following selection, dual-gene expression percentages in single-gene vector cotransduced and dual-gene vector transduced populations were similar. Equivalent

MGMT expression levels were obtained with each strategy, but GFP expression levels derived from the IRES dual-gene vector were significantly lower. In mice, selection of transduced cells resulted in equivalent vector-insertion averages for both the dual-gene vectors and the cotransduced single-gene vectors. These data demonstrate that cotransduction is an efficient alternative to dual-gene vectors and offers flexibility in

83 application, since previously-established selection conditions could be used for a range of therapeutic vectors.

Hypothesis

The percentage of dual-gene expressing cells enriched after cotransduction with single gene lentivectors will be similar to percentages obtained using dual-gene lentivectors.

Further, the level of dual-gene expression in cotransduced cells will exceed that of cells transduced with the dual-gene vectors.

Results

Comparison of dual-gene transfer strategies in vitro

Four self-inactivating lentiviral vectors were constructed; two single-gene vectors containing MGMT-P140K (pMGMT) or enhanced GFP (pGFP), and two dual-gene vectors containing either the EMCV-IRES (pMIG) or FMDV-2a (pMAG) elements to co- express MGMT and GFP (Fig. 29). The MGMT-P140K gene was placed in the 5' position of the dual-gene vectors to ensure that the level of drug resistance gene expression was equal to that obtained with the MGMT single-gene vector. Equivalent levels of expression and drug resistance were verified in K562 human erythroleukemia cells transduced with equal amounts of MGMT, MIG, or MAG virus (Fig. 30).

84 Figure 29: Lentivector constructs for comparison. Schematic diagram of the self- inactivating lentiviral vectors used to compare dual-gene transfer strategies. Each construct contains the MND promoter/enhancer, central polypurine tract and central termination sequence (cPPT/cts), as well as the woodchuck hepatitis virus PRE. The same 5' restriction site was used in the construction of each vector to ensure that all transcripts have the same 5' UTR sequence.

85 Figure 30: Drug resistance levels are equivalent with each strategy. Drug- resistant CFU survival percentages in K562 cells transduced with the MGMT single-gene vector, or the dual-gene MIG and MAG vectors.

As a first approach to comparing cotransduction with single-gene vectors to

transduction with the MIG and MAG dual-gene vectors, K562 cells were transduced

using the same total MOI for each method. As shown in figure 31, the percentage of

cells that expressed both MGMT and GFP after cotransduction was similar to that

obtained with the dual-gene MIG vector. This similarity was observed using MOIs of 1

or 5 for both sets of transductions and was maintained after selection with 10 μM BG and

25 μM BCNU. Since the total MOIs used for each transduction strategy were equivalent,

the MGMT and GFP gene equivalents used for the single-gene vector cotransductions

were a fraction of those used in the dual-gene vector transductions. However, efficient

expression from the cotransduced single-gene vectors and poor GFP expression from the

dual-gene IRES vector resulted in an equivalent dual-gene expression percentage with

86 Figure 31: Comparison of dual-gene transfer strategies in K562 cells. Total MGMT+ or MGMT-GFP dual+ percentages obtained in K562 cells with each dual-gene transfer strategy. Equivalent total MOIs (=1, or 5) were used for each transduction. Transduced cultures were untreated (UT) or treated (T) with 10 μM BG and 25 μM BCNU.

each strategy. Poor IRES-mediated translation efficiency reduced the percentage of dual- gene expressing cells detected (i.e., expression of MGMT was consistently greater than

GFP). MGMT and GFP were both efficiently expressed in K562 cells transduced with the dual-gene MAG vector and resulted in the highest percentage of dual-gene expressing cells in the untreated and drug-treated cultures. Interestingly, the GFP MFI in MAG- transduced cells was 3 to 8 times higher than in cotransduced or MIG-transduced cells, respectively (Fig. 32). Western analysis revealed that approximately half of the MGMT expressed from the dual-gene MAG vector was still fused to GFP, indicating that 2a hydrolase is not always effective in this construct (Fig. 33). However, this did not affect

DNA repair activity, as previously reported for an intentional MGMT-GFP fusion product (Choi et al. 2004). As expected, nuclear localization of MGMT co-localized

87 GFP to the nucleus (Fig. 34). Concentration of GFP in the nucleus increased the fluorescence intensity detected by flow cytometry. Therefore, MFI values for GFP were much higher in cells transduced with the dual-gene MAG vector, compared to cells transduced with the other dual-gene delivery strategies.

Figure 32: Relative MFI of MGMT or GFP in K562 cells. Relative MFI of MGMT or GFP in transduced K562 cell populations depicted in Figure 31 (MOI = 1). MGMT and GFP were measured separately by flow cytometry.

88 Figure 33: Dual-gene transfer expression levels. Western analysis of MGMT and GFP in protein extracts obtained from the drug-selected cultures in Fig. 31 (MOI = 1). The last lane in each blot is a control (K562 cells transduced with pMND-GFP-2a- HoxB4) that demonstrates a lack of fusion products when the same 2a sequence is used to link a different gene pair.

Figure 34: GFP localization after dual-gene transfer. K562 cells were transduced with each vector using a total MOI of 5 (2.5 of each vector for cotransduction). At 72 hrs, the cells were cytospun onto slides, fixed in 4% paraformaldehyde, and DAPI- stained for nuclear visualization. GFP and DAPI fluorescence were visualized separately and the fluorescent images were superimposed using Image J software. Due to fusion with MGMT, GFP localizes to the nucleus in MAG-transduced cells.

89 As a first approach to comparing the dual-gene delivery strategies in primary

cells, 5-Fluorouracil enriched murine BM-MNCs were transduced. Higher MOIs were

used to overcome the reduced transduction efficiency of primary murine cell populations.

For each dual-gene delivery strategy, BM-MNC cultures were transduced using a total

vector MOI of 20. Although the total MOI was held constant for each experiment,

staggered (5:15) MGMT:GFP vector MOI ratios were used for cotransductions since

these proportions were previously shown to result in the highest proportion of dual-gene

expressing cells after drug treatment. The total MGMT and dual-gene expression

percentages in the single-gene vector cotransduced BM-MNC cultures were lower, due to the lower individual vector MOIs used, and the interdependence of the MGMT and GFP

vector transduction rates (Fig. 35A). After selection with 20 μM BG and 25 μM BCNU, the percentage of dual-gene expressing cells in the single-gene vector cotransduced cultures (67%) was similar to that observed in cultures transduced with the dual-gene

MIG vector after drug treatment and culture expansion (62%). Drug treatment only led to moderate enrichment of MGMT-expressing cells in cultures transduced with the dual- gene vectors since these cultures had higher initial MGMT expression percentages (i.e., higher initial transduction resulted in less selection pressure).

90 A

B

Figure 35: Dual-gene transfer and expression in murine BM-MNCs. (A) Total MGMT or MGMT and GFP dual-gene expressing BM-MNC percentages obtained with each dual-gene transfer strategy. Equivalent total MOIs (= 20) were used for each strategy. Cotransductions were carried out using MGMT-GFP (M-G) vector MOI ratios that were equivalent (10-10) or staggered (5-15). Transduced cultures were untreated (UT) or treated (T) with 20 μM BG and 25 μM BCNU, and expanded in culture 10 days prior to analysis. (B) CFU survival percentages in transduced BM- MNC cultures following treatment with 20 μM BG and 0-25 μM BCNU.

91 The initially low MGMT expression percentages in the cotransduced cultures

resulted in fewer BM progenitor colony-forming units (CFU) surviving drug treatment, compared to cultures transduced with the dual-gene vectors (Fig. 35B). Nevertheless, dual-gene expressing cells enriched from the single-gene vector cotransduction cultures expressed both genes at high levels (Fig. 36). Equivalent MGMT expression levels were obtained in all of the BM-MNC cultures after drug treatment, but GFP expression levels in each cohort were significantly different (MAG > cotransduction, p < 0.001;

MAG > MIG, p < 0.0001; cotransduction > MIG, p < 0.01). As in K562 cells, GFP remained fused to MGMT and localized to the nucleus in BM-MNCs transduced with the

MAG dual-gene vector, resulting in elevated MFIs. The level of GFP expression in single-gene vector cotransduced cells was 50% higher than in cells transduced with the

MIG dual-gene vector. Thus, in primary hematopoietic cells, cotransduction is limited by lower initial dual-gene expression percentages, while IRES vectors are limited by poor

GFP expression levels and the 2a vectors are limited by inefficient protein processing.

Figure 36: Relative MGMT or GFP fluorescence in BM-MNCs. Relative MFI of MGMT or GFP expression in BM-MNCs depicted in Figure 35.

92 Transduction efficiencies are reduced in more primitive hematopoietic cell populations.

To determine whether the dual-gene expression percentages and the GFP expression

levels differed in more primitive populations, similar comparisons were carried out in

SKL cells. SKL cell cultures were transduced using a constant total vector MOI of 100

for each dual-gene delivery strategy. However, the individual MGMT and GFP single-

gene vector MOIs were staggered (20:80, MGMT:GFP) for cotransduction of the SKL

cells. Higher MOIs were required to efficiently transduce SKL cells, compared to BM-

MNCs. As an indicator of the overall lower transduction rate of SKL cells, the 5-fold higher dual-gene vector MOIs used to transduce SKL cells resulted in dual-gene expressing cell percentages that were similar to those obtained in BM-MNCs transduced with lower MOIs.

Transduction with the dual-gene MAG vector resulted in the highest percentage of dual-gene expressing cells, in both untreated and drug treated cultures (Fig. 37). The untreated SKL cell cultures cotransduced with the single-gene vectors had the lowest percentages of MGMT, GFP, and dual-gene expressing cells. After drug treatment, the

percentage of dual-gene expressing cells in the single-gene vector cotransduced cultures

(73%) was only slightly lower than that obtained with the dual-gene MIG vector (86%).

Since independent transduction events are required for cotransduction, the reduced transduction efficiency of SKL cells likely affected the dual-gene expression percentages in the cotransduction cultures to a greater extent than that in the dual-gene vector transduced cultures.

93 A

B

Figure 37: Dual-gene transfer and expression in murine SKL cells. (A) Total MGMT or MGMT and GFP dual-gene expressing BM-MNC percentages obtained with each dual-gene transfer strategy in SKL cells. Equivalent total MOIs (= 100) were used for each strategy. Cotransductions were carried out using MGMT-GFP (M-G) MOI ratios that were equivalent (50-50) or staggered (20-80). Transduced cultures were untreated (UT) or treated (T) with 20 μM BG and 25 μM BCNU, and expanded in culture 10 days prior to analysis. (B) CFU survival percentages in transduced SKL cultures following treatment with 20 μM BG and 0-25 μM BCNU.

94 As expected, the percentage of drug-resistant CFU in MIG and MAG transduced

cultures were equivalent at both BCNU doses evaluated, while CFU survival rates in cotransduced cultures were proportionately lower, since these cultures were exposed to lower MOIs of MGMT and the liquid outgrowth of surviving cells came from a smaller surviving progenitor population (Fig. 37B). Reflecting this greater degree of selection pressure in favor of gene expression, post drug-selected liquid cultures of cotransduced progenitors expressed GFP at levels twice as high as those transduced with the dual-gene

MIG vector (Fig. 38).

Figure 38: Relative MGMT or GFP fluorescence in SKL cells. Relative MFI of MGMT or GFP expression in SKL cell cultures depicted in Figure 37.

Comparison of dual-gene transfer strategies in vivo

To evaluate dual-gene expression over a longer time frame and after more robust expansion conditions, we compared each strategy in vivo. Donor SKL cell populations

were transduced using a total MOI of 100 for each strategy, which was staggered (20:80,

MGMT:GFP) for cells cotransduced with the single-gene vectors. Lethally irradiated

recipients were transplanted with 1000 SKL cells from each transduction culture and 2 x

95 106 lineage positive BM-MNCs for hematopoietic support. At 5 weeks post-transplant the mice were untreated or treated with three rounds of 30 mg/Kg BG and 10 mg/Kg

BCNU, allowing 3 weeks recovery between each treatment. MGMT and GFP expressing cell percentages were measured separately in PB-MNCs a day prior to each treatment.

Separate MGMT and GFP measurements were carried out since the expression levels were too low to resolve single and dual-gene expressing cell populations.

Figure 39: Dual-Gene expression in PB-MNCs after in vivo selection. Total MGMT or GFP expressing PB-MNC percentages 5 and 15 weeks after transplant. Animals were untreated (UT) or selected (3 X 30 mg/Kg BG and 10 mg/Kg BCNU), with 3 weeks between each round of selection. Numbers on the x-axis represent individual animals in cohorts transplanted with cotransduced (COTD), MIG- transduced, or MAG-transduced SKL cells. Values at 5 weeks are the expression percentages prior to the first treatment. Values depicted at 16 weeks are expression percentages in untreated (UT) or thrice-treated (3 x BG + BCNU) animals.

96 At 5 weeks post-transplant, but before drug selection, the percentage of MGMT- expressing PB-MNCs were equivalent (5 ± 4%) in each cohort (Fig. 39). GFP- expressing PB-MNC percentages in cotransduced and MAG-transduced cell recipients were similar (14 ± 5% and 13 ± 4%, respectively), while MIG-transduced cell recipients had a lower percentage of GFP-expressing PB-MNCs (5 ± 3%). After drug treatment and

recovery, the average percentages of MGMT-expressing and GFP-expressing PB-MNCs

were enriched to a similar degree in each cohort; the average MGMT-expressing PB-

MNC percentage in cotransduced recipient animals increased 11-fold (from 5± 3% to 58±

22%), compared to 8-fold increases (from 5± 4% to 37± 24%) in MIG-transduced cell

recipients, and 9-fold increases (from 5± 2% to 47± 17%) in MAG-transduced cell

recipients. Average GFP-expressing PB-MNC percentages increased by 7-fold (from 5±

3% to 34± 18%) in MIG-transduced cell recipients, and by 5-fold (from 13± 4% to 65±

15%) in MAG-transduced cell recipients. Enrichment for GFP-expressing cells was

lower in the cohort cotransduced with the single-gene vectors (3-fold; from 14± 5% to

39± 8%) due to expansion of MGMT singly-transduced cells and the loss of GFP singly-

transduced cells. One animal in the MIG-transduced cohort died after the third treatment.

MGMT-expressing PB-MNCs in this animal declined after the first and second

treatments, indicating a lack of engraftment with drug-resistant cells.

The animals were sacrificed and BM-MNCs were harvested at 17 weeks post-

transplant. At this point, stable hematopoiesis post-drug selection allowed analysis of progenitor selection as impacted by the different transduction strategies. Drug selection increased the proportion of gene expressing cells (Fig. 40). MGMT-expressing BM-

97 MNC percentages in single-gene vector cotransduced cell recipient animals increased from 0-14% in the untreated controls to an average of 77 ± 27% in the drug treated animals. In the same cohort, the percentage of GFP-expressing BM-MNCs increased from 5± 4% in untreated animals to 61 ± 29% in the drug treated animals. Among the dual-gene vector cohorts, the average MGMT-expressing and GFP-expressing BM-MNC percentages in MIG-transduced recipients were not significantly different between drug treated and untreated control animals (MGMT, p = 0.075; GFP p = 0.105 ) since one of the untreated animals engrafted with a high percentage of dual-gene expressing cells in the absence of selection. In contrast, in the MAG-transduced cell recipients, MGMT- expressing BM-MNC percentages increased from 4± 2% in untreated controls to 84 ± 3% in the drug treated animals. Likewise, GFP-expressing BM-MNC percentages increased from 4± 2% to 79 ± 18% after drug treatment. In addition multilineage selection was observed, with increased GFP-expression in B220+ and Gr-1+ BM-MNC fractions (Fig.

41).

98 Figure 40: Dual-gene expression in BM-MNCs after in vivo selection. Total MGMT and GFP expressing BM-MNC percentages derived from transplant recipients depicted in Figure 39.

Figure 41: Dual-gene expression in lymphoid and myeloid populations. BM- MNCs derived from untreated (UT) or drug-selected (3 x BG + BCNU) animals were labeled with B220 (B-cells) or Gr-1 (granulocytes) antibodies for detection of GFP expression in lymphoid or myeloid lineages, respectively. Animal numbers correspond to those in figs. 39-40.

99 Real-time PCR analysis was used to determine the total number of viral insertions

per cell in CFU recovered from each cohort (Fig. 42). While the proportion of

transduced CFU increased after drug treatment, the average number of insertions in marked CFU was not significantly different among the different cohorts of drug treated

animals. In drug-treated animals, the number of vector insertions per marked cell averaged 5 ± 3 (range 1.7 ± 0.2 to 10 ± 1) in cotransduced recipients, compared to 6 ± 5

(range 1.3 ± 0.5 to 14.1 ± 0.1) and 4 ± 1 (range 2.6 ± 0.1 to 6 ± 1) in MIG-transduced and

MAG-transduced recipients, respectively. These data indicate that: 1) selection does not skew towards cells containing high copy numbers, and 2) that cotransduction with separate single-gene vectors does not result in an overall increase in vector insertions, compared to dual-gene vector strategies. As with the in vitro experiments, the level of

MGMT expressed in MGMT-positive BM-MNCs was not significantly different among

the cohorts, while the GFP expression levels were (Fig. 43). As expected, GFP

expression was localized to the nuclei of MAG-transduced cells, which would be

problematic in therapeutic applications in which the gene products were required in other

cellular compartments. GFP expression levels in cotransduced recipient BM-MNC cells

was over 3-fold higher than that in cells transduced with the MIG vector (p < 0.0005).

Thus, cotransduction with single-gene MGMT and therapeutic vectors should result in

higher therapeutic gene expression levels after drug selection, compared to dual-gene

IRES vectors.

100 Figure 42: Vector insertion averages are equivalent with each strategy. The average number of vector insertions per cell, calculated from individual CFU. Animal numbers correspond to figs. 39-41. Values were determined using primers for both MGMT and GFP on each CFU (n = 8-15 CFU/animal). Mean MGMT and GFP values were calculated from all of the CFU analyzed in each animal. Mean MGMT and GFP values were totaled (cotransduced CFU), or averaged (MIG and MAG transduced CFU) to get the total number of vector insertion per cell.

We estimated the numerical expansion based on engrafted cell transduction and expression rates, and on the total cellular reconstitution levels. On this basis, and on the assumption that the normal mouse bone marrow compartment consists of 3 x 108 MNCs, dual-gene expressing cells in cotransduced recipient animals were enriched by an average of 1.8 ± 0.5 x 107-fold, compared to 1.8 ± 0.9 x 107-fold in MIG-transduced recipients

and 1.18 ± 0.08 x 106-fold in MAG-transduced recipients. The level of MAG-transduced

cell expansion was lower due to lower selection pressure, since higher dual-gene

expressing cell percentages were present in the transduced donor cell populations. Thus, the overall level of dual-gene expressing cell enrichment in the cotransduced and MIG- transduced recipients were equivalent.

101 These data indicate that, after drug selection, similar levels of vector insertions are obtained with each strategy. In addition, the single-gene vector cotransduced populations contained sufficient numbers of early progenitors to engraft multilineage

cells expressing both genes at high levels. Thus, cotransduction with single-gene vectors

may be an efficient alternative to dual-gene vectors for coupling hematopoietic

progenitor selection to therapeutic gene expression, and it avoids the need to customize

new dual-gene therapeutic vectors that co-express mutant MGMT.

Figure 43: Relative MGMT and GFP MFI in BM-MNCs. Relative MFI of MGMT or GFP expression (measured separately) in the BM-MNCs of each transplant recipient cohort depicted in figure 40.

102 DISCUSSION:

Studies aimed at coupling selective stem cell enrichment to therapeutic gene

expression have focused on the use of dual-gene vectors. Many strategies have been

employed to co-express two genes from the same vector, including the use of alternative

splice sites, internal promoters, gene fusions, IRES elements, and ribosome slippage sites.

However, higher MOIs are often needed for dual-gene vectors to compensate for low expression levels. The work reported in this study shows that MGMT-mediated selection

can be used to efficiently enrich hematopoietic progenitor cells cotransduced with

separate single-gene lentivectors. Although the initial dual-gene transfer rates are

reduced with this strategy, the percentage of dual-expressing cells that are enriched after

drug selection are similar to percentages obtained after transduction with dual-gene

vectors. In addition, the populations cotransduced with the single-gene vectors expressed

both gene products at high levels with vector insertion averages equivalent to those

obtained with dual-gene vectors.

Advantages and problems exist for each dual-gene vector strategy that are often

specific to the particular application or gene combinations used. Strong selective

pressure applied to alternative splicing or internal promoter strategies can lead to preferential splice site recognition or promoter interference, favoring drug resistance gene

expression at the cost of therapeutic gene expression (Emerman et al. 1984). Gene

fusion is likely the most efficient strategy, but requires that both gene products are

functional as a fusion and active in the same cellular compartment. Recently Amendola

et al. reported the creation of dual-gene lentivectors using bidirectional promoters

(Amendola et al. 2005). These promoters co-express two genes that are positioned in

103 opposite orientations to one another in the same vector. However, the efficiency of this

strategy in the setting of stringent drug selection remains to be determined.

The most common dual-gene expression strategies utilize IRES or ribosome

slippage sequences for co-expression. Therefore, we compared cotransduction of single-

gene MGMT and GFP vectors to dual-gene vectors that utilize IRES or ribosome

slippage sites to co-express MGMT and GFP. The EMCV and FMDV elements were

specifically chosen, since these sequences have been thoroughly characterized and are

routinely used in dual-gene expression vectors. The dual-gene MIG vector used in these studies was constructed to maximize IRES-mediated GFP expression. IRES activity has been reported to be most efficient when the IRES element is placed less than 100 nucleotides downstream of the first gene (Attal et al. 1999). Therefore, the IRES in the dual-gene MIG vector was inserted 70 nucleotides downstream of MGMT. The position of the second gene's start codon with respect to the EMCV-IRES sequence is critical, since a specific AUG site is preferentially used for IRES-initiated translation (Davies et

al. 1992). Thus, GFP was cloned into an Nco I site that was engineered to position the

GFP start codon at the optimal site for IRES-initiated translation. Although the same

gene pair had to be maintained for a comparative study, cis-acting complications have

also been attributed to the specific genes used in bicistronic constructs (Hennecke et al.

2001). In the studies reported here, MIG-transduced cells efficiently expressed MGMT, but the GFP expression levels were 2 to 3-fold lower than the levels obtained in cells cotransduced with the single-gene vectors, despite similar vector insertion averages. Low

IRES-mediated translation efficiencies have been reported using other gene combinations, and IRES elements (Hennecke et al. 2001). IRES-mediated expression

104 appears to be particularly low in lentiviral vectors, but this is likely due to the lower transcription rates obtained with internal lentivector promoters, compared to LTR- mediated transcription in retroviral vectors.

The FMDV-2a element has also been thoroughly characterized and successfully used in dual-gene vectors (Ryan et al. 1994; de Felipe et al. 1999; Furler et al. 2001;

Milsom et al. 2004; Lengler et al. 2005). The linkers used to join MGMT and GFP to the

2a consensus have been used in our lab to efficiently co-express other gene combinations.

While, the C-terminal 2a fusion residues abrogated gene function in one such construct, no evidence of fusion products were previously detected. Similar complications have been reported by others. Lengler et al. found that the C-terminal 2a residues specifically destabilized some gene products and not others (Lengler et al. 2005). Recently Milsom et al. reported the use of a triple-gene vector, containing an MGMT-2a-GFP cassette similar to that of the dual-gene MAG vector, reported here (Milsom et al. 2004). Although slightly different linker sequences were used, similar levels of MGMT-GFP fusion products were reported. Thus, the efficiency of 2a processing appears to be more related to the specific gene combinations used, rather than a cell type-specific phenomenon.

Since MGMT repair activity was unaffected by fusion with GFP, the most efficient dual gene expression strategy in the studies reported here was achieved with the dual-gene

MAG vector.

Unlike dual-gene vector strategies, which may be limited by lower dual-gene expression levels or inefficient protein processing, cells cotransduced with the single- gene vectors expressed both genes at high levels. Cotransduction is limited by the reduced dual-gene transfer rates inherent to this strategy. However, MGMT-mediated

105 selection efficiently enriched limiting numbers of cotransduced progenitor cells. Lower

MGMT gene equivalents were used in the single-gene vector cotransduction experiments to keep the total MOIs equivalent. This led to lower initial percentages of drug-resistant cells, compared to the percentages obtained with the dual-gene vectors. However, after selection the percentage of MGMT-expressing cells in the cotransduced cultures were equivalent to the percentages obtained with the dual-gene vectors. Thus, the level of expansion did not appear to be a limiting factor for cotransduction with the single-gene vectors.

While the cotransduction studies reported here relied on the individual vector transduction rates, strategies that physically couple the separate vectors should improve the efficiency of cotransduction. Zhang et al. reported that low molecular weight polylysine polymers allow low-speed concentration of VSVG-pseudotyped lentivectors

(Zhang et al. 2001). The virus-polylysine precipitates in these studies resembled "beads on a string." Thus, a similar strategy might be used to co-precipitate the separate viruses used for cotransduction, using vector MOI ratios that are optimal for each application.

These studies demonstrate that cotransduction with single-gene vectors is an efficient alternative to dual-gene vectors for selective expansion of dual-gene expressing hematopoietic progenitor cells in vivo. Although higher levels of expansion are required with this strategy, MGMT-mediated enrichment was not limiting. Further, the single- gene vectors were less prone to the cis-acting complications in the dual-gene vectors that reduced protein expression or processing. Since the degree of selective expansion and expression obtained using the single-gene MGMT vector is independent from that of another cotransduced vector, previously-established MGMT vector MOIs and drug-

106 selection conditions could be applied to a wide range of therapeutic vectors, avoiding the need to repeat this process and construct separate dual-gene vectors for each application.

Materials and Methods

Vectors: A self-inactivating lentiviral luciferase vector, pCSO-rre-cppt-MCU3-LUC, containing the MND promoter/enhancer sequences (Myeloproliferative sarcoma virus

enhancer, Negative control region deleted, dl587rev primer-binding site substituted), the

rev response element (rre) and the cPPT/CTS was obtained from Donald Kohn.

Luciferase was removed from this vector by partial restriction with Nco I and Eco RI and

replaced with a multiple cloning site. The MGMT-P140K and enhanced GFP coding

sequences were separately inserted into the Nco I site (closest to the MND promoter) and

a unique Bam HI site. These constructs were then restricted at unique Bam HI and Eco

RI sites for insertion of the woodchuck hepatitis virus posttranscriptional regulatory

element (wPRE, provided by Thomas Hope, University of Illinois at Chicago) to generate

the pMND-MGMT and pMND-GFP single-gene vectors. The encephalomyocarditis virus IRES element was obtained from the pCITE-2a vector (Clontech). The IRES sequence was deleted from pCITE-2a by partial Pvu II and Msc I restriction and the resulting vector fragment was religated. 5'-NotI and 3'-EcoRI sites were added onto the

IRES sequence by PCR and the restricted PCR product was added back to the IRES- deleted CITE vector, generating an IRES vector with both 5' and 3' multiple cloning sites, pSOB. GFP was cloned into pSOB using Nco I and Bam HI to position the GFP start codon at the site reported to maximize IRES-mediated translation (Davies et al. 1992).

The IRES-GFP cassette was then removed and placed 70 bp downstream of MGMT

107 using a unique Not I site in pMND-MGMT, generating pMND-MIG. The FMDV

sequence was generated from oligos using overlap extension PCR. An MGMT-2a PCR

fragment was generated using the oligos, MGMT-FOR; 5'-gtgagcagggtctgcacgaa-3' and

MGMT-2A; 5'- ctgccaacttgagcaggtcaaagctcaaaagctgtttcaccggtgc gtttcggccagcag-3'. A

separate 2a-GFP fragment was generated using the oligos 2A-GFP; 5'-

ggcaggggacgtcgagtccaaccctgggcctatggtgagcaagggcga-3' and GFP-REV; 5'-

ctagagcggccgctttacttgtacagctcgtcc-3'. The Resulting MGMT-2a and 2a-GFP fragments

were annealed, extended, and amplified with the MGMT-FOR and GFP-REV oligos.

The resulting fragment was cloned into pMND-MIG using a unique Sfi I site within

MGMT and a Not I site 3' of GFP, generating pMND-MAG. All constructs were verified

by sequencing.

Virus preparation and transductions: Virus was generated as described by Zielske et

al.(Zielske et al. 2003) Briefly, 293T cells were co-transfected with the packaging vector

(pCMVdeltaR8.91), the VSVG pseudotyping vector (pMD.G), and either the pMND-

transducing vectors at a mass ratio of 3:1:3, using Lipofectamine 2000 (Invitrogen)

according to the manufacturer's instructions. Virus produced 24-48 hours after transfection was harvested in Dulbecco’s modified Eagle medium (DMEM; Cellgro) containing 10% heat-inactivated FBS (Cellgro) and 2 mM GlutaMAX (Invitrogen).

Virus-enriched media was filtered through 0.45 μm syringe filter units (Millipore) and stored at -80° C. Expression titers were determined on K562 cells using virus dilutions that resulted in less than 10% transduction. Titers ranged from 0.5-3 x 107 expression units/ml. Virus preps used for cotransductions were premixed at the specified MOI

108 ratios prior to the addition of cells. K562 cells were cotransduced in Iscove’s medium

(Cellgro) containing 10% heat-inactivated FBS, 2 mM GlutaMAX, and 8 μg/ml polybrene (Sigma, St. Louis, MO). Whole BM and sorted stem cell populations were transduced for 12 hours in alpha-MEM containing 20% heat-inactivated FBS, 2 mM

GlutaMAX, 6 μg/ml polybrene, in the presence of 20 ng/ml murine IL-3, and 50 ng/ml of murine IL-6, and SCF (R&D Systems Inc., Minneapolis, Minnesota). MOIs defined for mouse cell transductions were based on expression titers established in K562 cells.

Transplants: SKL cells were isolated as previously described (Reese et al. 2003).

Transduced cells were transplanted immediately after transduction. Six-week-old recipient mice were lethally irradiated with 850 cGy using a Cs137 source and transplanted with 1,000 cells from each transduction culture, along with 2 x106 lineage+ cells per mouse for hematopoietic support. The animals were sacrificed at 17 weeks for expression and vector insertion number analysis in BM-MNCs and CFU, respectively.

Drug Selection: BG was synthesized by Dr Robert Moschel at the Frederick Cancer

Research Institute (Frederick, MD). BCNU was obtained from the Drug Synthesis and

Chemistry Branch of the National Cancer Institute (NCI; Bethesda, MD). All in vitro drug treatment incubations were carried out in serum free media at 37° C. Cells were pretreated with BG for 1 hour, followed by a 2 hour treatment in BCNU. K562 cells were treated in Iscove's media. Drug treatments for in vitro murine BM selections were carried out with 25 μM BG and 0, 5, 15, or 25 μM BCNU in alpha-MEM containing

1.2% spleen cell conditioned media (Stem Cell Technologies). Drug treated murine BM

109 samples were expanded 10 days in liquid culture using the same media formulation as

described for transduction, without Polybrene. Progenitor drug treatment survival assays were performed as previously described.(Davis et al. 1997) For in vivo selection, BG was dissolved to 3 mg/ml in 40% polyethylene glycol (Union Carbide Corp., Danbury,

Connecticut) and 60% PBS (pH 8.0), and BCNU was dissolved in ethanol and diluted to

1 mg/ml in PBS. Mice were injected intraperitoneally with 30 mg/kg BG, followed by 10 mg/kg BCNU an hour later. Three rounds of selection were carried out at 3 week intervals, beginning at 5 weeks.

Flow cytometry: All flow cytometry results were obtained with an LSR flow cytometer

(Becton Dickinson). GFP expression was measured in K562 cells after a single wash in

PBS. For MGMT detection, K562 cells were fixed in 2% paraformaldehyde for 30 minutes at 4° C, permeablized in 1% Tween-20 for 30 (K562 cells) or 60 (murine cells) minutes at 37° C, and blocked in 10% normal goat serum for 15 minutes at room temperature. The cells were then immunolabeled with the mouse anti-human MGMT monoclonal antibody MT3.1 (Kamiya Biomedical, Seattle, WA) and an (APC)- conjugated goat anti-mouse secondary antibody (Caltag Laboratories, Burlingame, CA).

Dual detection of MGMT and GFP expression was carried out with the same staining protocol as that used for MGMT alone. Red blood cells were lysed in murine PB and

BM samples prior to preparation for flow cytometry.

Copy number analysis: The number of vector insertions per cell was evaluated using a

LightCycler instrument and the LightCycler FastStart DNA Master SYBR Green I kit

110 from Roche Diagnostics (Basel, Switzerland). MGMT and GFP copy numbers were determined using proviral-specific MGMT primers; Sense 5'-

ACGTCTATATCATGGCCG-3', Antisense 5'-TGAGGATCTTGACAGGATT-3', or

GFP primers; Sense 5'-ACGTCTATATCATGGCCG-3', and Antisense 5'-

TGTGATCGCGCTTCTC-3'. Genomic copy numbers were determined using GAPDL4-

specific primers, as previously described.(Zielske et al. 2003) MGMT, GFP, and

GAPDL4 copy numbers were assessed using a standard curve consisting of murine

genomic PB-MNC DNA spiked with known MAG vector copy numbers, and verified

using K562 cell clones containing a single MAG vector insertion. One diploid copy of

the mouse genome was assumed to contain two copies of GAPDL4 and 5.9 pg of DNA.

MGMT, GFP, and GAPDL4 copy numbers were analyzed in 8-15 CFU were per animal.

The average number of vector insertions per marked CFU in cotransduced cells is the

cumulative average of the total MGMT and GFP copies, whereas the average number of

vector insertions per marked CFU in dual-gene vector transduced cells is the combined

average of the individual MGMT and GFP copy numbers.

111 CHAPTER 5

SCREENING RETROVIRUS LIBRARIES FOR MUTANTS

WITH EXPANDED HOST RANGE

Summary

Most of the viral envelope proteins used for pseudotyping gene transfer vectors utilize ubiquitously-expressed receptors. Although these pseudotypes allow a particular delivery system to be used in a broad range of therapeutic applications, the lack of target cell specificity results in an increased risk of insertional mutagenesis. Improving vector affinity towards therapeutically relevant cell types should increase the transduction efficiencies of target cells and reduce the level of integration events in non-target cell populations. Viral binding specificities have been improved by inserting targeting motifs into envelope proteins but these insertions often disrupt the subsequent envelope fusion activity required for viral entry. We set out to improve virus specificity using a combinatorial strategy and an entry-based screen to identify virions with expanded tropism. The receptor binding domain of the ecotropic murine leukemia virus was randomized by PCR and the resulting envelope library was cloned into a retroviral vector containing an IRES-GFP cassette. This vector library was packaged into VSVG- pseudotyped virions to generate a virus library, which was then transduced (via VSVG) into cell lines that are resistant to ecotropic virus infection. Gag-pol expressing cells lines were used to ensure that envelope mutations allowing transduction of non- permissive cells would be packaged and amplified in culture. Although the proof of principle was demonstrated by the amplification of virions in permissive cell populations,

112 no mutants with an expanded host range were identified. These data indicate that larger libraries may be needed to identify mutants with an altered host range.

Hypothesis

The receptor binding domain of the ecotropic envelope protein can be randomized to create a library of mutant virions displaying unique envelope proteins and specific virions within this mutant pool will have an expanded tropism that includes human cells.

Further, characterization of the envelope sequences present in early, intermediate, and late stages of amplification, in cells expressing the wild-type receptor, will allow the positions randomized to be ranked by functional importance.

Background

The murine leukemia virus (MuLV) family is one of the most widely studied groups of retroviruses and has been used extensively in the construction of gene transfer vectors. These viruses have been classified into five subgroups based on their host range and interference patterns: ecotropic, xenotropic, polytropic, amphotropic, and 10A1

(Miller 1996). Ecotropic MuLVs bind to a cationic amino acid transporter (Rec1) present only on murine and rat cells (Kim et al. 1991; Wang et al. 1991), whereas xenotropic and polytropic MuLVs bind to a different set of receptors expressed on a variety of species

(Rein et al. 1984; Rein 1990). The amphotropic MuLV subgroup recognizes a highly conserved phosphate symporter (Ram1/Pit2) expressed on most mammalian cells, and the

10A1 retrovirus is capable of infecting cells through two different receptors, the gibbon ape leukemia virus receptor (GALVR/Pit1) and the amphotropic receptor (Ram1/Pit2)

113 (Kavanaugh et al. 1994; Miller et al. 1994; van Zeijl et al. 1994). Interestingly, only three amino acid changes in the amphotropic envelope protein are required to allow this virus to utilize the GALVR/Pit1 (Han et al. 1997). Although each subgroup utilizes a different receptor for infection, the sequences encoding their SU subunits are highly conserved. Sequence alignments have revealed four regions within SU (VRA, VRB,

VRC, and PRO) that vary between the different MuLV subgroups and are involved in receptor specificity (Battini 1995; Fass et al. 1997; Han et al. 1998). Structural alignments of the amphotropic and 10A1 envelope sequences were carried out using the

Friend ecotropic envelope structure as a template. The alignments illustrate the highly conserved scaffolding regions in the center of each envelope protein, with the variable regions radiating outward and devoid of significant secondary structure (Fig. 44).

Point mutations have been used to map the residues in the ecotropic envelope protein that are responsible for receptor interactions (MacKrell et al. 1996; Bae et al.

1997; Davey et al. 1999; Zavorotinskaya et al. 1999). Davey et al. demonstrated that

Arg83, Glu86, and Glu87 are important residues for receptor binding, and identified Asp84 as a crucial residue for binding and infection (Davey et al. 1999). Fass et al. resolved the crystal structure of the Friend ecotropic envelope protein receptor binding domain (Fass et al. 1997). Based on this and the point mutant information, they proposed that the receptor-interacting region of this envelope consists of a charged ridge and a hydrophobic pocket (Fig. 45). The ecotropic envelope receptor (mCAT-1) shares 87% identity to the human protein (hCat-1), with only two amino acid differences responsible for human cell resistance to ecotropic virus infection (Albritton et al. 1993). Therefore, randomization

114 of the residues responsible for the ecotropic envelope protein's receptor specificity may broaden the host range of this virus to include human cells.

Figure 44: Structural alignment of MuLV envelope receptor-binding domains. The protein sequences of the Friend ecotropic, 10A1, and amphotropic envelope protein receptor binding domains were aligned using Swiss-PdbViewer. The aligned sequences were structurally aligned with Swiss-Model, using the known structure of the Friend ecotropic RBD as a modeling template (Fass, D. et al. 1997). A central "scaffolding" region (gray) was maintained within each RBD, with variable regions (colored) radiating from into the periphery.

115 90

83

Figure 45: The ecotropic envelope receptor binding domain. MuLV envelope proteins are highly conserved. Three variable regions (A, B, and C) have been identified in the ecotropic virus. Key residues within the "charged ridge" of variable region A are essential for receptor binding.

Results

To evaluate whether randomization of the ecotropic envelope protein would

generate virions with an expanded host range, a library screen was established using transduction as a basis for selection. PCR-based randomization was used to generate a library of ecotropic mutants encompassing 5 residues in the receptor binding domain; three reported to be involved in binding (Arg83, Glu86, and Glu87), one that is required for

infection (Asp84), and one uncharacterized residue (Thr90) (Fig. 46).

116 Randomization Oligos

83 84 86 87 90

Sequencing Chromatogram of Randomized Vector Library

VECTOR

Figure 46: Randomization of the ecotropic envelope gene for retroviral display. C An oligo with 5 randomized codons (NN /G) was used to create a mutant ecotropic envelope library. The five codons randomized are located in the “charged ridge” of the SU subunit. Four of the five randomized residues (Arg83, Asp84, Glu86, and Glu87) have been shown to be involved in virus-receptor interactions. The randomized PCR product contained unique restriction sites for in-frame insertion into MSCV-EIG. The resulting MSCV-EIG library was sequenced to verify randomization of the target codons.

The randomized PCR product was inserted into a retroviral vector containing the ecotropic envelope and an IRES-GFP cassette (MSCV-EIG). The resulting vector library, representing 3.3 x 106 of 4 x 106 possible variants, was co-transfected along with a vector expressing the VSVG envelope into an env-deficient packaging cell line to generate a VSVG-pseudotyped virus library (Fig. 47). The virus library was screened in cell lines that express the retroviral gag and pol genes, but are resistant to ecotropic virus infection. Therefore, cells transduced with the VSVG-pseudotyped library produced virions pseudotyped with a specific mutant ecotropic envelope protein and the

117 corresponding mutant gene packaged within. Mutants capable of subsequent transductions should deliver the successful env gene allowing production and amplification in culture. Since GFP is transcriptionally linked to env expression, horizontal amplification was assessed by increases in the percentage of GFP-expressing cells over time.

Figure 47: Production of retroviral envelope libraries. The MSCV-EIG vector library was cotransfected with a VSVG expression vector into 293 cells stably expressing the retroviral gag and pol genes (Phoenix-GP). The virus-enriched supernatant was then used to transduce packaging cell lines at a low density. The transduced cells were then evaluated for increased GFP expression percentages, indicative of horizontal transfer of mutants conferring entry function. Alternatively, GFP expression percentages were evaluated over prolonged periods after transfection.

118 The horizontal virus amplification and GFP detection screen was first evaluated

by transfecting permissive (AM12) and non-permissive (E86) packaging cells with the

MSCV-EIG vector expressing the wild-type ecotropic envelope. Transient GFP expression in E86 cells diminished with each passage since these cells already express the ecotropic envelope and are therefore protected from superinfection (Fig. 48).

Figure 48: Viral tropism and receptor interference in packaging cell lines. Packaging cell lines that express a particular envelope protein resist superinfection from virus baring the same type of envelope. Ecotropic virus will infect the murine AM12 and NIH 3T3 cell lines, but not human cells. Ecotropic envelope expression in E86 cells prevents transduction with ecotropic virus. Amphotropic virus will infect murine and human cells that do not express the amphotropic envelope.

However, the percentage of GFP-expressing AM12 cells increased with each passage.

The initial percentage of GFP-expressing cells was 5% at two days post-transfection, which increased to 100% by day 9 (data not shown). To test the feasibility of limiting amounts of permissive virus to be amplified in culture, AM12 cells were transduced with the MSCV-EIG vector, or a control vector expressing CXCR4 in place of the envelope

119 gene. GFP-expressing cells were then diluted into a 1 x 105-fold excess of untransduced

AM12 cells. After 13 days of culture in the presence of Polybrene a slight increase in the percentage of GFP-expressing cells was detected in cultures that received the MSCV-EIG

transduced cells (<0.5% to 1%), which peaked at 97% after an additional 9 days in

culture (Fig. 49). GFP expression was undetected in cultures spiked with control-

transduced cells. As shown in figure 50, the GFP-expressing cell percentages increased

exponentially. From this, the GFP expression percentages (~ transduced cells) were

calculated to double every 19 ± 0.5 hours. Thus, we assumed that mutant virions would

have time to amplify in cells that were passaged (1:6) on every third day.

Figure 49: Env-expressing retroviral vectors are amplified in gag+-pol+ cells. The MSCV-EIG vector, or a control (CXCR4-IRES-GFP) vector, were transfected with VSVG into Phoenix-GP cells. The virus-enriched supernatant collected at 48 hours post-trasnfection was used to transduce the amphotropic AM12 packaging cell line. GFP-expressing cells were then diluted into a 1 x 105-fold excess of untransduced AM12 cells and GFP expressing cell percentages were evaluated every 3 days by flow cytometry.

120 Figure 50: Exponential transfer of an Env-expressing RV vector in gag+-pol+ cells. Log-linear analysis of amplified GFP expression in AM12 cells. The linearity of the plot indicates that amplification of GFP in AM12 cells is exponential. Thus, the doubling time for GFP expression was calculated to be 19 ± 0.5 hrs. (ln2/(0.38±.01)).

Screening the retroviral library in gag+-pol+ cell lines

Non-permissive human cell lines (Phoenix-GP, or TEFLY) or the murine E86 packaging cell line were transduced to screen the virus library. All of the transduced cell lines expressed the retroviral gag-pol genes to ensure that successful mutants would be amplified in culture. AM12 cells were transduced with the library to evaluate the length of time required to amplify mutants utilizing the wild-type receptor. As shown in figure

51, screening results in the E86 cell line indicated that some of the mutants were able to enter these cells through an alternative pathway. The vector expressing the wild type envelope protein did not result in an increased percentage of GFP-expressing cells, and

121 no replication competent virus was detected when the supernatant was transferred to

NIH3T3 cells. Supernatant collected from the library-transduced E86 cells at passage 11 was able to transduce naïve E86 cells, but no further amplification was observed. These results suggest that the observed amplification likely occurred from reduced expression of the ecotropic envelope in untransduced cells, reducing their protection from superinfection.

6 9 15 18 21 24 27 30 33 36 39 42 DAYS

Figure 51: Transduction-based screens in E86 cells. Supernatant containing the VSVG-pseudotyped virus library (E86 LIB), or virus baring the wild-type ecotropic envelope (E86 WT) was used to transduce the murine E86 packaging cell line to detect gain of function mutants that permitted entry in the setting of receptor interference. GFP expressing E86 percentages were evaluated over time by flow cytometry. Virus mutants allowing entry through a novel receptor should be amplified in culture over time. Approximately 4.8 x 105 library variants were screened.

Library screens in the Phoenix-GP packaging cell line indicated that none of the mutants were capable of entering human cells, or that the screen was not carried out long

122 enough for amplification to occur (Fig. 52). The gradual decline in the percentage of

GFP-expressing cells may have resulted from promoter silencing or a growth advantage in favor of untransduced cells. The percentage of GFP-expressing cells also decreased in

Phoenix-GP cultures transduced with the control MSCV-EIG vector expressing the wild- type envelope gene.

6 9 15 18 21 24 27 30 33 DAYS

Figure 52: Transduction-based screens in 293 cells. Supernatant containing the VSVG-pseudotyped virus library (LIB1-4), or virus bearing the wild-type ecotropic envelope (EIG WT) was used to transduce the human gag-pol expressing cell line (Phoenix-GP) to detect gain of function mutants that permitted entry. Virus mutants allowing entry through a novel receptor should be amplified in culture over time. Approximately 2.4 x 106 library variants were screened.

In AM12 cells, the control vector only took 16 days to spread to over 90% of the cells, while amplification in the library screens were first detected by day 15 and reached 90%

123 of the population by day 33 (Fig. 53). Therefore, gain of function mutations in non-

permissive cell screens should have been detected in the time frames used.

3 6 9 12 15 18 21 24 27 30 33 36 39 DAYS Figure 53: Transduction-based screens in AM12 cells. Supernatant containing the VSVG-pseudotyped virus library (AM12 EIG LIB1-3), or virus bearing the wild-type ecotropic envelope (AM12 EIG WT) was used to transduce the murine AM12 amphotropic packaging cell line to detect mutants that retained entry function through the natural receptor. Mutants maintaining wild type receptor binding should be amplified in culture over time. Approximately 1.2 x 106 library variants were screened in the library plates.

To increase the size of the library represented in each experiment, the vector

libraries were transfected into each of the gag-pol expressing cell lines. As in

transduction-based AM12 screens, the control MSCV-EIG vector spread rapidly in culture, reaching 100% of the cells by 5 days in culture (Fig. 54). Library clones that

expressed functional envelope proteins also spread to 100% of the cells, but required an

additional 32 days of culture.

124 Figure 54: Transfection-based screens in AM12 cells. The envelope vector library virus library (ENV LIB1-3), a vector expressing the wild-type ecotropic envelope (WT ENV), or a control retroviral GFP expression vector (MFG-GFP) were transfected into AM12 cells to detect mutants that retained entry function through the natural receptor.

No viral mutants capable of spreading in E86 cells were detected (Fig. 55). A slight increase in the percentage of GFP-expressing E86 cells was evident in cultures transduced with the wild-type control vector, likely due to a loss of stable envelope expression (i.e., resistance to superinfection). The transfection-based screens did not result in detectable amplification in human Phoenix-GP cells (data not shown).

Therefore, the vector library was transfected into another human packaging cell line

(TEFLY), derived from human fibrosarcoma cells (Fig. 56). As in the Phoenix-GP cell screens, the GFP-expressing TEFLY percentages declined rapidly after transfection.

However, a slight increase (<.1% to ~0.3%) in the percentage of GFP-expressing TEFLY cells was detected in one of the library screen cultures at day 25, while the percentage of

125 GFP-expression in other cultures continued to decline. The GFP-positive cells were sorted and culture expanded. Media collected from these cells was added to naive

TEFLY cells, but no subsequent transmission of the GFP expression vector was detected.

Thus, the low level of GFP expression detected in the initial culture screen was likely due to stably transfected cells, rather than spread of a functional virus clone.

Figure 55: Transfection-based screens in E86 cells. The envelope vector library virus library (ENV LIB1-3), a vector expressing the wild-type ecotropic envelope (WT ENV), or a control retroviral GFP expression vector (MFG-GFP) were transfected into E86 cells to detect gain of function mutants that permitted entry in the setting of receptor interference.

126 Figure 56: Transfection-based screens in TEFLY cells. The envelope vector library virus library (ENV LIB1-3), a vector expressing the wild-type ecotropic envelope (WT ENV), or a control retroviral GFP expression vector (MFG-GFP) were transfected into the human gag-pol expressing fibrosarcoma cell line (TEFLY) to detect gain of function mutants that permitted entry.

127 Discussion

Although the entire library was screened in human packaging cell lines, no

mutants with an expanded tropism were detected. The screens in non-permissive cell

lines were carried longer than the time required to amplify mutants utilizing the wild-type

receptor (i.e., screens in AM12 cells). Mutants utilizing unique receptors may have been

diluted out during cell passage due to reduced binding or entry rates. However, the cells

were only split 1:6 every three days, which should have allowed sufficient time for

productive transduction of neighboring cells. Additional screens, in which cells exposed

to the virus library are sorted for GFP expression, may lead to the identification of

mutants with an expanded host range.

Since these studies were carried out, other groups have identified hydrophobic

residues, including Trp100 and Trp142, that appear to confirm the role of the hydrophobic

pocket for additional receptor interactions (Davey et al. 1999; Zavorotinskaya et al.

1999). These studies indicate that randomization of other residues, in addition to those in the charged ridge, may be required for this strategy to work. However, in the primary sequence the hydrophobic residues are distant from those in the charged ridge, which would complicate any targeted randomization strategies involving PCR-based techniques.

Davey et al. have demonstrated that single mutations that had no effect on binding or entry resulted in significantly reduced infectivity when combined as double mutants

(Davey et al. 1999). This suggests that randomization of only a few key sites may increase the likelihood of finding functional mutations, rather than including sites that could be involved in stabilizing the protein. Alternatively, other point mutations have been identified that enhance fusogenicity or prevent insertion-mediated destabilization of

128 the envelope (Zhu et al. 1998; Zavorotinskaya et al. 1999). Thus, randomization of the receptor binding residues in envelope proteins that already contain the fusogenic or stabilizing mutations may increase the chances of finding gain of function mutations.

The fact that MuLV envelope proteins all recognize ubiquitous transporter proteins as receptors makes it unlikely that this randomization strategy would lead to the identification of a virus mutant that recognizes a specific cell type. More realistically, it would lead to a mutant which recognized another ubiquitous-expressed receptor. Thus, at the very least, we expected to amplify mutants able to utilize human CAT-1, since only two amino acid differences provide ecotropic resistance to human cells.

Despite the inability to identify mutants with an expanded host range, these studies should provide valuable information about the importance of the specific sites randomized. An undergraduate student, Mourad Ismail, is currently studying the significance of the specific sites randomized by evaluating how fast the particular residues converge to the wild-type sequence. This is being done by sequencing individual virions in the mutant population at early, intermediate, and late stages of amplification. The rate at which virions with specific "wild-type" residues are amplified should provide insight into the functional importance of the specific sites randomized.

Materials and Methods

Vectors. The retroviral vector, MSCVneoEB, was obtained from Dr. Robert Hawley

(George Washington University Medical Center, Washington DC). The neomycin gene was removed from MSCVneoEB by restriction with Bgl II and Bam HI and replaced with an IRES-GFP cassette from pSOB-GFP (described previously) to generate MSCV-IG.

129 An ecotropic envelope expression vector, pENV, was obtained from Arthur Bank at

Columbia University. The envelope gene was removed from pENV by Eco RI restriction

and cloned into a unique Eco RI site of MSCV-IG located 5' of the IRES-GFP cassette.

The resulting vector was designated MSCV-EIG. Sac II and Bam HI sites were removed

from MSCV-EIG by fill-in reactions with the Klenow fragment of DNA polymerase I.

Sac II and Stu I sites were incorporated into the VRA region of the ecotropic envelope using the oligos, ecoriENV3-for; 5'-ggcgccggaattcgtggaccatcctctagactgac-3' and stuIENV-rev; 5'-aggcctgtggggcccggggca-3', and the Sac II site was generated using the oligos, sac2ENV11-for; 5'-gcggtgcaacactgcctggaa-3' and ENVid4-rev; 5'- tttgggcttggagagtggct-3'. The resulting PCR fragments were linked using SOE PCR

(Horton et al. 1989) and the product was inserted into MSCV-EIG using unique Eco RI and Bam HI sites.

Envelope randomization. Residues in the receptor binding domain of VRA (R83, D84,

E86, E87, T90) were randomized with the forward ecoriENV3 oligo (above) and the reverse oligo, RAND2; 5'- caacgtggcgccccactccctNN(G/C)atttccNN(G/C)NN(G/C)cgtNN(G/C)NN(G/C)ccttgtcggac ccgacgac-3'. Randomization was achieved by doping with A, T, C, and G in the first two nucleotide positions, and doping with C or G at the last nucleotide position in each codon. Individual nucleotide concentrations in the doping pools were adjusted to ensure equal nucleotide incorporation rates. The randomized PCR products were restricted with

Sac II and Stu I and ligated into MSCV-EIG. The ligation products were then electroporated into DH5α to generate a vector library consisting of 4 x 106 variants.

130

Cell lines. Murine NIH 3T3-based packaging cell lines, GP+E86 and GP+AM12, were provided by Arthur Bank (Columbia University). Packaging cell lines derived from the human 293 cell lines (Phoenix-Eco, Phoenix-Eco, and Phoenix-GP) were obtained from

ATCC. The gag-pol expressing human fibro sarcoma cell line, TEFLY, was obtained from Dr. FL. Cosset (Lyon, France). All of the gag-pol, or packaging cell lines were maintained in DMEM containing 10% heat-inactivated FBS and 2 mM GlutaMAX.

Library screens. For transduction-based screens, a VSVG-pseudotyped virus library was generated by co-transfecting a VSVG expression plasmid (pMD.G) and the MSCV-EIG library into Phoenix-GP (gag+-pol+) cells. After 12 hours, the transfection complex was replaced with DMEM containing 10% heat-inactivated FBS and 2 mM GlutaMAX.

Virus-enriched supernatant was harvested 48 hours after the start of transfection and titered on the human HEK-293T cell line. Titers ranged from 1-5 x 105 GFP expression units/ml. Virus-enriched media was added to target cells in the presence of 8 μg/ml

Polybrene. Transduction efficiency was assessed at 48 hours using flow cytometric measurement of the GFP expression percentages. The remaining cells were split into multiple plates to evaluate horizontal spread of the vector in culture. Each plate was then split 1:3 to 1:6 (depending on the growth rates of each cell line) every third day and replated in media containing Polybrene. GFP expression percentages were evaluated in the remaining cells at each passage. For transfection-based screens, the randomized

MSCV-EIG vector library was transfected into each cell line using Lipofectamine 2000, according to the manufacturer's recommendations. GFP expression percentages were

131 evaluated 36 hours after transfection and the remaining cells were diluted with

untransfected to bring the GFP expression percentages to levels that could be evaluated

for increases or decreases. As with transduction-based screens, the transfected cells were split 1:3 or 1:6 every third day, evaluated by flow cytometry, and replated in media containing Polybrene.

132 CHAPTER 6

UTILIZING PHAGE PEPTIDE LIBRARIES FOR TARGETED

TRANSDUCTION STRATEGIES

Summary

Insertion of a binding domain into a viral envelope protein is a commonly used strategy

for engineering targeted vectors. However, even with successful processing, viral

display, and target-cell binding, most of these strategies fail at the step of cell entry. We

previously attempted to use randomized envelopes in a retroviral display strategy to

redirect viral tropism. However, the library sizes represented were limited by subcloning

efficiencies and by the titers that can be attained with retroviral vectors. Filamentous

bacteriophage are commonly used to display vast peptide libraries that can be panned

over target molecules to identify binding sequences. To take advantage of the peptide diversity represented within phage libraries for display in retroviral vectors, we panned phage libraries over HSC-enriched murine BM-MNC populations. Thus, the phage population would be restricted in size but biased for the intended target cell population.

The aim of this project was to then display the binding-peptide population within insertion-tolerant regions of the ecotropic envelope protein to screen for viral clones that

preferentially transduce target cells. This chapter describes the consensus sequences

obtained from phage panning experiments and the further characterization of specific

phage clones or peptides to identify their targets. The vector constructs and strategies for

utilizing phage peptides for targeted transduction are described in the following chapter.

133 Hypothesis

Diverse phage peptide libraries can be panned over SKL cell populations to reduce the

library to populations that both fit within the constraints of retroviral display strategies,

and are biased for SKL-cell binding. Further, display of these "SKL-binding" peptides in an insertion-tolerant region of the ecotropic envelope protein will improve viral affinity to SKL cells, without perturbing envelope function.

Results

Panning phage peptide libraries for SKL -binding sequences

Two strategies were used to enrich phage populations with affinity towards stem cells (Fig. 57). In the first strategy, a phage peptide library was panned over whole BM-

MNC populations. Stem cells were isolated from the panned BM-MNCs population using fluorescence activated cell sorting for the SKL cell fraction (~ 0.05% of the total cells). This strategy was based on the assumption that phage clones that bind ubiquitously-expressed surface proteins would be subtractively removed from the library due to the vast excess of lineage-positive cells in whole bone marrow. In the first round of panning 2 x 1011 plaque-forming units (PFU) of library phage were added to 23 x 106

BM-MNCs (Table 1). Approximately 9,600 SKL cells were isolated, from which 5,300

PFU were obtained. This process was repeated three times, with the output phage at each

round used as input for the subsequent panning experiment. After the third round of panning 5 phage clones were sequenced, all of which had the consensus sequence

(designated "SKL").

134 Figure 57: Panning phage libraries for stem cell binding epitopes. A disulfide constrained heptapeptide phage library was added to whole murine bone marrow cells (A) or infused via tail vein into mice (B). Phage were eluted off of sorted SKL cell populations and amplified. Individual phage clones were sequenced and characterized after each round of panning.

135 Table 1. Summary of In Vitro SKL-Cell Panning Experiments.

INPUT CELLS SKL CELLS TOTAL PFU CLONES PAN PFU PANNED ISOLATED RECOVERED SEQUENCED 1 2 X 1011 23 X 106 9671 5340 0 2 7 X 1010 40 X 106 6011 - 0 3 2 X 1010 50 X 106 2977 2670 5 = WTLDRGY

The second strategy used was based on the report by Pasqualini et al. that randomized phage libraries could be used to identify in vivo homing peptides (Pasqualini et al. 1996). The same heptapeptide phage library (2 x 1011 PFU) was diluted in PBS and infused into the tail vein of one mouse. After 10 minutes the BM-MNCs were extracted and labeled for SKL cell sorting. Phage eluted from the sorted SKL cell populations were then amplified and sequenced. Three rounds of panning were carried out (Table 2).

For the third round of panning two animals were infused with 2 x 1011 PFU of the output phage from the second round of panning. However, one mouse was sacrificed after 10 minutes, and the other mouse was sacrificed after 30 minutes. Interestingly, different phage sequences were eluted from the SKL cells isolated from each mouse (Table3).

Table 2. Summary of In Vivo SKL-Cell Panning Experiments.

INPUT TIME TO SKL CELLS TOTAL PFU CLONES PAN PFU SACRIFICE (min) ISOLATED RECOVERED SEQUENCED 1 3.6 X 1011 10 8112 89 8 2 4 X 1011 10 7004 - 8 3A 2 X 1011 10 2000 22 8 3B 2 X 1011 30 3500 156 8

Although the SKL consensus sequence identified during the in vitro panning experiments was also identified among the in vivo panning sequences, no consensus sequence was identified. As shown in table 3, individual clones were present as rare sequences in multiple pans, or as multiple sequences in one particular pan. In all 15 different

136 sequences were identified. Thus, phage that home or bind to rare hematopoietic cells can

be isolated and enriched from a vast excess of heterogeneous cell populations.

Table 3. Phage Clones Isolated from In Vivo SKL-Cell Panning. CLONE SEQUENCE PAN 1* PAN 2* PAN 3A* PAN 3B** A withheld X2 B withheld X1 X1 X1 C withheld X1 D withheld X1 E withheld X1 F withheld X1 X2 X1 G withheld X1 H withheld X1 I withheld X2 X4 J withheld X1 K withheld X4 L withheld X2 M withheld X1 N withheld X1 SKL withheld X1 X2 *10 minute post-infusion BM harvest **30 minute post-infusion BM harvest

Characterization of SKL cell-binding clones

Blast searches using the identified sequences were mostly uninformative, due to the limiting number of residues in the query sequences. Individual clone preparations were biotinylated to evaluate the degree of binding to BM-MNC populations. Cell-bound phage were labeled with APC-conjugated streptavidin and detected by flow cytometry.

As shown in figure 58, the majority of phage clone only bound to a small percentage

(<0.2%) of BM-MNCs. However, 5% of the BM-MNCs were labeled using the SKL

137 Figure 58: Phage clone binding to murine BM-MNCs. The phage clones identified in SKL cell panning experiments were biotinylated and added to murine BM-MNCs. Cell-bound phage were labeled with APC-conjugated avidin and detected by flow cytometry. Since free amines were biotinylated, clones with lysine residues in their consensus sequences may have been inactivated (Lysine residues are denoted with asterisks). phage clone. A similar experiment was carried out to determine the specificity of binding. The individual biotinylated phage clones were added to BM-MNCs in the presence of a 10-fold excess of non-biotinylated wild-type phage, or an equivalent amount of the corresponding, non-biotinylated phage. Many of the clones remained bound to a fraction of BM-MNCs in the presence of the wild-type phage competitor, but the specific competitor reduced the degree of binding. The lowest binding and competition percentages were observed in clones that contained lysine residues in their recognition sequence. Thus, the binding activity was likely disrupted by biotinylation.

138 Figure 59: Phage BM-MNC binding in the presence of unlabeled competitor. Biotinylated phage clone preps were added to murine BM-MNCs in the presence of a 10-fold excess of unlabeled wild-type phage, or an equivalent amount of the same, unlabeled phage. The biotinylated phage were labeled with APC-conjugated avidin and analyzed by flow cytometry. Since free amines were biotinylated, clones with lysine residues in their consensus sequences may have been inactivated (Lysine residues are denoted with asterisks).

The low overall binding percentages to whole BM-MNC fractions suggested that the phage clones recognized rare cell populations. However, identification of the specific cell populations each clone recognized was hampered by variation in biotinylation reactions and the high level of nonspecific binding. Therefore, two peptides were synthesized for further characterization; one based on the SKL-clone sequence, and the other based on the M-clone sequence. Clone M was only represented as one sequence in the third round of panning, however, this sequence showed a high degree of homology to

139 the rat robo-1 gene. Robo-1 homology was interesting due to its role in CXCR4- mediated homing of neural crest progenitors and breast cancer cells (Prasad et al. 2004;

Jia et al. 2005). Both peptides were synthesized to include a biotinylated lysine residue just prior to the C-terminus (Fig. 60). The peptides were then used to screen BM-MNC fractions co-labeled with antibodies that recognize specific hematopoietic lineages.

Figure 60: Cyclic peptides synthesized. Cyclic peptides were synthesized, based on the consensus sequences of two phage clones isolated from SKL cell panning. The "SKL" peptide sequence is derived from the clone found in both the in vitro and in vivo panning experiments.

The SKL peptide co-labeled the majority of the CD11b-positive PB-MNCs and BM-

MNCs (Fig. 61A), but only a fraction of the B220-positive populations. CD11b (αM chain of Mac-1 integrin) is primarily expressed on granulocytes and macrophages, but can also be expressed in dendritic cells, natural killer cells, and microglia. However, the

SKL-labeled cells localized to forward- and side-scatter populations characteristic of granulocytes. The SKL peptide was subsequently found to primarily bind granulocytes

(Gran1-positive).

140 A.

B.

Figure 61: Titration of the SKL peptide on murine PB and BM MNC fractions. Murine PB and BM MNC fractions were labeled with PE-cojugated antibodies that recognize distinct hematopoietic cell lineages. (A) SKL-peptide titration on CD11b+ PB-MNCs and BM-MNCs. (B) SKL-peptide titration on B220+ PB-MNCs and BM- MNCs. 2 x 105 cells per sample were exposed to 0.5-10 μg of the cyclic SKL peptide. Bound peptide was labeled with APC-streptavidin and detected by flow cytometry. The percentage listed in each box represents the percent positive in the total cell population pictured.

141 The clone M peptide was similarly screened using several hematopoietic cell lineage markers to determine if it recognized a specific bone marrow cell fraction.

Interestingly, clone M bound to approximately 1% of the total BM-MNC population.

The bound population represented 30% of the TER119+-erythroblasts (Fig. 62).

Unfortunately, the CFU frequency in lineage-negative BM-MNC populations labeled with either peptide was equivalent to that of the lineage-negative BM-MNC population alone. Thus, these peptides are not associated with the progenitor cell phenotype.

Although neither of the peptides characterized were shown to specifically bind to progenitor cells, these studies demonstrate that diverse phage libraries can be minimized to phage pools that target distinct cell populations. These enriched phage pools can then be evaluated in function-based screens (discussed in chapter 7).

Figure 62: Murine BM-MNC fractions bound by cyclic peptides. Murine BM- MNCs were labeled with the cyclic peptides were stained with a panel of antibodies recognizing different murine hematopoietic cell surface antigens. Co-labeled cell fractions were detected by flow cytometry.

142 Screening phage peptide libraries for virus-binding sequences

Phage libraries were also panned over env-expressing cell lines, or virion-coated plates, in attempts at identifying peptide sequences that specifically bind to viral envelope proteins. Nonspecific phage were removed from the library populations using subtractive pans over envelope-deficient packaging cells (Phoenix-GP cells). The unbound phage population was then added to Phoenix cells that expressed either the ecotropic or amphotropic envelope proteins. For the second round of panning, NIH-3T3 cells were used for subtractive phage panning, and E86 (ecotropic) or AM12 (amphotropic) cells for positive phage panning. The third and fourth rounds of panning were carried out using virus-coated plates. Similar experiments were carried out to identify VSVG-binding peptides. However, the toxicity of VSVG required the use of transiently transfected cells for the positive panning experiments.

Sequences identified in the third and fourth pans are listed in table 4.

No VSVG-binding consensus sequences were identified. However, four of the seven ecotropic-env clones encoded the same peptide sequenc. Of the four clones with the same peptide consensus, two different DNA sequences were found to encode it, suggesting specificity to the target. Five amphotropic-env clones were sequenced after the third round of panning and three distinct peptide sequences were identified. Only two different sequences were in the five clones analyzed after the fourth round of panning.

Three of the four peptide sequences identified in the amphotropic-env phage clones were highly similar, but the low number of clones sequenced and the small size of the peptides displayed prevented the identification of a definitive consensus sequence.

143 Table 4. Phage Clones Isolated from Envelope Panning. TARGET ENV SEQUENCE PAN3* PAN4* ECOTROPIC withheld 1/5 0/2 ECOTROPIC withheld 1/5 0/2 ECOTROPIC withheld 3/5 1/2 ECOTROPIC withheld 0/5 1/2 AMPHOTROPIC withheld 2/5 0/5 AMPHOTROPIC withheld 1/5 0/5 AMPHOTROPIC withheld 2/5 3/5 AMPHOTROPIC withheld 0/5 2/5 VSV-G withheld 1/5 0/5 VSV-G withheld 2/5 0/5 VSV-G withheld 1/5 0/5 VSV-G withheld 1/5 0/5 VSV-G withheld 0/5 1/5 VSV-G withheld 0/5 1/5 VSV-G withheld 0/5 1/5 VSV-G withheld 0/5 1/5 VSV-G withheld 0/5 1/5

* Number of clones with listed sequence / Total clones sequenced

Discussion

These data reiterate previous findings that phage libraries can be used to identify binding interactions. However, we set out with the novel objective of using these

libraries, in conjunction with fluorescence-activated cell sorting, to identify peptides that

home or bind to surface proteins found on stem cell populations. The fact that the two

peptides characterized recognize BM-MNC populations that are co-labeled with lineage-

positive antibodies suggests that the screens were not carried out under conditions in

which the lineage-positive cells were saturated with antibody. Longer incubations after

in vivo infusion of the phage particles may increase the likelihood of finding homing

144 peptide sequences, since HSC homing is thought to require longer time frames (Szilvassy

et al. 2001). However, incubations longer than 30 minutes led to the death of one mouse,

presumably due to anaphylactic shock from contaminating levels of LPS in the phage

preps.

Recently, another group used this technique to isolate a peptide sequence that

homes to the murine bone marrow compartment (Nowakowski et al. 2004). However,

evidence that the peptide sequence was involved in homing was based on the slower

engraftment of cells pre-incubated the entire phage, while control cells were not exposed

to phage. The objective of this study was to enrich binding peptides for subsequent

entry-based screens. A detailed description for futures studies utilizing these peptides is

presented in the following chapter.

Materials and Methods

SKL-cell panning experiments. A disulfide-constrained heptapeptide phage library was purchased from New England Biolabs (Ipswich, MA). This library represents 1.2 x 109

transformants at 1 x 1010 PFU/μl. For in vitro panning experiments, whole bone marrow

cells were isolated from the femurs and tibia of C57 black mice, and the red blood cells

were lysed in ammonium chloride lysis buffer. Cells and the phage library (2 x 1011

PFU) were blocked separately in PBS containing 4% (w/v) non-fat dry milk (NFDM), 1

mM PMSF, 20μg/ml aprotinin, and 1μg/ml leupeptin. After blocking, the phage library

was added to the cells and rotated at 4° C for 2-3 hours. The cells were then washed in

FACS buffer (PBS + 2% BSA), and PE-conjugated antibodies (B220, CD3, CD4,

Ter119, andCD11b) were added to mark the differentiated cell populations. The lineage-

145 positive fraction was then depleted using anti-PE conjugated magnetic beads, and the

remaining cells were labeled with APC-CD117 (c-Kit) and FITC-Sca1 antibodies. SKL

cell fractions were then isolate using a FACSARIA fluorescence activated cell sorter (BD

Biosciences, San Jose, California). Sorted SKL cells were centrifuged and suspended in

200 μl of PBS. Phage were eluted from the cells by adding 300 μl of 76 mM citric acid

(pH 2.5) and incubating 5 minutes at room temperature. The cells were centrifuged and

the phage eluate was neutralized in 400 μl of 1 M Tris-HCl (pH 7.5). The phage eluate

was then added to a log-phase culture of E. coli (ER2738) and amplified according to the

manufacturer's protocol. A 10 μl aliquot of unamplified phage eluate was set aside to

determine the raw titer of the sorted phage population. Phage amplified from each round

of panning were used as input for subsequent panning experiments. For in vivo panning

experiments, the phage were diluted in PBS and up to 2 x 1011 PFUs were infused into

the tail vein of C57 black mice. The mice were sacrificed 10-30 minutes later and the

phage bound to the SKL cell fraction were isolated, as above. Phage amplifications and

preparations were carried out according to the manufacturer's recommendations.

SKL-cell panning experiments. All cells and phage preps were blocked in PBS containing 4% (w/v) non-fat dry milk (NFDM), 1 mM PMSF, 20μg/ml aprotinin, and

1μg/ml leupeptin before mixing. For the first round of panning, the phage library (2 x

1011 PFU) was added to Phoenix-GP (env-deficient) packaging cells and incubated for 2-

3 hrs., rotating at 4° C. The cells were removed by centrifugation and the remaining phage were added to env-positive (Phoenix-Eco, or Phoenix-Ampho) packaging cell lines. After incubation, the cells were washed and the bound phage were eluted, as

146 described for the SKL-cell panning experiments. This process was repeated for the

second round of panning, but NIH-3T3 cells were used for subtractive pans, and E86, or

AM12 cells were used for ecotropic or amphotropic envelope panning, respectively. The

final panning rounds were carried out in plastic dishes preloaded with virus-enriched

media (serum-free). Phage amplifications and preparations were carried out according to

the manufacturer's recommendations.

Biotinylation of phage clones. Approximately 8 x 1010 PFUs from a clonal phage prep

were suspended in 300 μl of a 100 mM NaHCO3 (pH 8.6) solution containing 10 μg/ml

Biotinamidocaproic acid 3-sulfo-N-hydroxy succinate ester. The biotinylation reaction

was incubated at 25° C for 1 hour and terminated by adding 10 μl of 5 M NH4Cl.

Biotinylated phage were then precipitated with 100 μl of PEG-NaCl (20% (w/v)

polyethylene glycol–8000, 2.5 M NaCl), suspended in PBS and stored at 4°C.

Screening biotinylated phage preps for binding to murine BM-MNC populations.

Clonal preps of biotinylated phage were each added to 1 x 106 murine BM-MNCs (~1000

PFU/ cell) and rotated at 4° C for 1 hour. The cells were washed in FACS buffer, and 1

μg of APC-conjugated streptavidin (CALTAG laboratories, Burlingame, CA) and PE-

conjugated lineage antibodies were added. After a 30-minute incubation at 4° C, the cells were washed in FACS buffer and analyzed using a Becton Dickinson LSR1 flow cytometer. Binding competitions were carried out for specific biotinylated phage clones by preloading cells with an equivalent amount of unlabeled clone, or a 10-fold excess of unlabeled wild-type phage.

147

Cyclic peptide screens. Fourteen-residue peptides, representing two SKL-binding consensus sequences, were synthesized on an Omega 396 multiple peptide synthesizer

(Advanced ChemTech, Louisville, KY). The peptides were synthesized to include a biotinylated lysine residue at position 13, and the products were oxidized to generate the disulfide bridge resembling the configuration of the phage-displayed peptides. The peptide products were suspended in PBS and stored at -80° C. The peptides were added to murine BM-MNC fractions, labeled with PE-conjugated lineage marker antibodies, and the bound peptides were detected using APC-streptavidin, as above.

148 CHAPTER 7

FINAL COMMENTS

HSC gene transfer vectors

Gene therapy has the potential for curing a wide range of inherited or acquired

hematopoietic disorders, including hemoglobinopathies, and immunodeficiency syndromes. Stable gene transfer into HSCs has improved with the use of lentiviral vectors, and efficient in vivo enrichment of transduced cells has now been demonstrated in both small and large animal models (Davis et al. 2000; Neff et al. 2003; Zielske et al.

2003). However, limited titers and HSC marking levels obtained with MuLV-based vectors led to an emphasis on gene transfer rate as a main indicator of vector efficacy.

Thus, MOI was often overlooked when higher transgene-expressing HSC percentages were observed with lentiviral vectors. Unfortunately, the insertional mutagenesis risk associated with integrating viral vectors has now been demonstrated in clinical gene therapy trials (Kaiser 2003). These findings gave cause to reevaluate the drive for high gene-transfer rates, and to redefine vector expression efficiency in terms of integrated copy numbers. Detailed characterization of lentivirus and retrovirus insertion site preferences have been reported (Schmidt et al. 2003), and additional studies are being carried out to determine the relative risks involved in using these vectors.

Despite the higher rate of gene transfer into HSCs with lentiviral vectors, the expression levels obtained with these vectors are often lower than those obtained with retroviral expression vectors (An et al. 2000). However, lentiviral vectors have been in development a short time compared to their retroviral counterparts and more work is

149 needed for optimal expression in hematopoietic tissues. HSC-optimized promoters

derived from retroviral vectors have been used as internal promoters in lentiviral

constructs (Richard et al. 2004). These promoters have increased lentivirus expression

levels in HSCs, but appear to be less active in the lentiviral backbone. In our early selection studies the lower lentivector-mediated expression levels were found to have a significant effect on the ability to enrich MGMT-transduced murine HSCs in vivo; in non-lethally irradiated recipient animals MGMT-expressing cells were only enriched to

10% of the total BM-MNC fraction after three rounds of BG and BCNU treatment.

Similar transplant experiments using 10-fold less MFG-MGMT transduced donor cells resulted in a 10-fold higher percentage of MGMT-expressing cells after selection (Davis et al. 2000). This may explain why reports of corrected murine hematopoietic disorders using MGMT and therapeutic gene lentivectors are limited to lethally-irradiated recipient animals. Thus, compared to retroviral vectors, the titers and expression-based transduction rates obtained with lentiviral vectors are higher, but the expression levels in murine hematopoietic cells are reduced. Additional studies should be carried out to compare retroviral and lentiviral gene expression levels (normalizing to copy number) in human hematopoietic progenitor cells.

150 Differential transduction rates of murine progenitor cell populations

Since higher levels of gene transfer were observed using lentiviral vectors and higher expression levels were obtained in cells transduced with single-gene lentivectors, we decided to evaluate the use of separate single-gene vectors to cotransduce cells. We hypothesized that inclusion of MGMT as one of the single-gene vectors would allow efficient enrichment of dual-gene expressing cells after drug treatment. In vitro studies using human cell lines demonstrated that the percentage of dual-gene expressing cells was directly proportional to the total percentage of cells expressing each single-gene vector. However, in primary murine cells, the percentage of dual-gene expressing cells was higher than expected, based on the individual MGMT and GFP expressing cell percentages. Higher MOIs were required to efficiently transduce murine BM-MNC populations. Thus, in mixed cell populations the cells that are susceptible to transduction were likely to be transduced with both single-gene vectors, while resistant cells would be resistant to both. Sutton et al. demonstrated that human hematopoietic progenitors in

G1/S/G2/M phases of the cell cycle were efficiently transduced with lentiviral vectors, while those in G0 were resistant (Sutton et al. 1999). In addition, the transduction efficiency of G0 cells was not increased by subsequent cytokine stimulation. Our in vitro and in vivo single-gene cotransduction experiments in BM-MNC and SKL cell populations were mostly carried out using the addition of a single dose of virus. Thus, cells in the proper phase of the cell cycle were likely transduced, while many of the early progenitor cells may have been in G0. The concentration of stem cells in 5-FU enriched

BM-MNC populations is reduced, compared to SKL cell fractions, but 5-FU treatment is likely to increase the percentage of actively cycling stem cells. This could explain why

151 few of the animals transplanted with transduced BM-MNCs engrafted with gene- expressing cells, but those that did were more likely to express both genes.

Long-term reconstituting stem cells have decreased expression of the c-kit

receptor and increased expression of CD11b after 5-FU treatment (Randall et al. 1997).

Therefore, 5-FU treatment is not carried out prior to SKL cell sorting. To minimize the

time for ex vivo manipulation of SKL cells, single-gene vector cotransductions were

carried out over relatively short time periods (~12 hours) prior to transplant. Thus, only

the cells cycling within the 12-hour time frame were likely to be transduced. This may

have led to higher short-term progenitor cell transduction rates and lower gene transfer

into more primitive stem cell populations. This could explain the reduced ability to

selectively enrich the percentages of dual-gene expressing cells to levels much above

those in animals transplanted with the transduced SKL cell fractions; limiting numbers of

dual-gene transduced stem cells and excessive numbers of dual-gene expressing

progenitors should effectively reduce selective expansion at the stem cell level. These

studies suggest that longer transduction periods, with multiple low-dose virus additions,

should increase the percentage of early progenitor cells transduced, compared to rapid

transduction protocols with higher MOIs. This extended transduction protocol should

increase the size of the susceptible cell population, thereby reducing the number of

integration events in a given cell.

152 Dual-gene transfer strategies

Comparative studies between dual- and single-gene vectors based on the ability to

transduce and enrich dual-gene expressing cells in vivo demonstrated the problems and advantages associated with each strategy. Cotransduction with single-gene vectors is mainly limited by the transduction efficiency of the target cell population. However, the cells that take in and express both vectors do so at high levels. The cotransduction strategy reduces the need for dual-gene vector construction, and allows the application of previously-established drug selection conditions to be applied to different therapeutic gene vector applications. This strategy is also advantageous in that it should permit stem cell selection to be coupled to high levels of lineage-specific therapeutic gene expression.

Since the two genes are co-delivered in dual-gene vector constructs, these vectors are more efficient at dual-gene delivery. However, several complications may arise using

these vectors that are often specific to the particular gene combinations used. Thus, co-

expression via dual-gene vectors often requires extensive trouble-shooting and the

evaluation of many different dual-gene linkage elements. Even after this process the

dual-gene expression levels are often reduced or complicated by other issues.

IRES-mediated translation levels were found to be limiting in retroviral vectors.

Therefore, we expected even lower levels of IRES-mediated translation with lentiviral

constructs, based on lower lentivector-mediated transcription rates. However, the

translation efficiency of an IRES element is often significantly impacted by its local

genetic environment; the specific gene pair used and their position will often determine

the efficiency of this element. IRES-mediated translation efficiency of GFP was

particularly low in murine BM-MNCs transduced with the MND-MIG vector. Other

153 dual-gene EMCV-IRES lentivectors constructed in our lab have also proved to be

inefficient. Several other IRES elements have been identified that may work better in the

context of lentiviral vectors and MGMT-mediated HSC enrichment. However, the time

required for these selection experiments often exceeded the time in which new IRES

elements were identified. Future studies could include additional IRES elements in an

attempt at identifying any that are particularly suited for this application.

Dual-gene expression efficiency using the FMDV-2a element appears to be

significantly dependent on the specific genes used and on the therapeutic application. In some constructs observed efficient 2a processing, but the 2a residues that remained on the first gene product abrogated its activity. In other constructs, such as the MND-MAG vector, the 2a processing efficiency was reduced but it had no impact on the activity of the gene products. Thus, the 2a element efficiency for dual-gene expression will only be determined by trial and error.

The cotransduction results reported in the current studies were based on straight mixtures of virus-enriched media. Thus, the percentage of dual-gene expressing cells

was determined by the individual vector transduction rates. Cotransduction rates might

be enhanced using targeted delivery vectors or molecular conjugates that physically link

virus particles. Fibronectin fragments increase HSC transduction rates by co-localizing

virus particles to the cells (Moritz et al. 1996). If multiple virus-binding domains exist in fibronectin fragments, these molecules might increase the rate of cotransduction with single-gene vectors. Similarly, Zhang et al. demonstrated that virions precipitated with poly-L-lysine resembled virus "beads" on a poly-lysine "string"(Zhang et al. 2001).

154 Thus, poly-lysine may be useful for precipitating mixtures of single-gene virus particles for improved dual-gene transfer rates.

CONCLUSION

The studies described in this dissertation demonstrate that single-gene lentiviral vectors can efficiently cotransduce murine progenitor cell populations, from which dual- gene expressing cells can be enriched with drug selection. They further demonstrate that cotransduction is an efficient alternative to dual-gene expression vectors, and can be used to link MGMT-mediated progenitor cell selection to a second, lineage-specific expression vector. Cotransduction with single-gene vectors avoids the necessity for generating dual- gene vectors and selection conditions for each application. Further, single-gene vectors are less prone to many of the expression issues that limit the effectiveness of dual-gene vectors. Results from additional studies are described that set the framework for future experiments aimed at targeting virus to specific cell types.

155 BIBLIOGRAPHY

Abbott, B. L. (2003). "ABCG2 (BCRP) expression in normal and malignant hematopoietic cells." Hematol Oncol 21(3): 115-30.

Abonour, R., D. A. Williams, L. Einhorn, K. M. Hall, J. Chen, J. Coffman, et al. (2000). "Efficient retrovirus-mediated transfer of the multidrug resistance 1 gene into autologous human long-term repopulating hematopoietic stem cells." Nat Med 6(6): 652-8.

Agarwal, M., T. W. Austin, F. Morel, J. Chen, E. Bohnlein and I. Plavec (1998). "Scaffold attachment region-mediated enhancement of retroviral vector expression in primary T cells." J Virol 72(5): 3720-8.

Ager, S., B. H. Nilson, F. J. Morling, K. W. Peng, F. L. Cosset and S. J. Russell (1996). "Retroviral display of antibody fragments; interdomain spacing strongly influences vector infectivity." Hum Gene Ther 7(17): 2157-2164.

Albritton, L. M., J. W. Kim, L. Tseng and J. M. Cunningham (1993). "Envelope-binding domain in the cationic amino acid transporter determines the host range of ecotropic murine retroviruses." J Virol 67(4): 2091-6.

Allay, J. A., B. M. Davis and S. L. Gerson (1997). "Human alkyltransferase-transduced murine myeloid progenitors are enriched in vivo by BCNU treatment of transplanted mice." Exp Hematol 25(10): 1069-76.

Allay, J. A., L. L. Dumenco, O. N. Koc, L. Liu and S. L. Gerson (1995). "Retroviral transduction and expression of the human alkyltransferase cDNA provides nitrosourea resistance to hematopoietic cells." Blood 85(11): 3342-51.

Allay, J. A., O. N. Koc, B. M. Davis and S. L. Gerson (1996). "Retroviral-mediated gene transduction of human alkyltransferase complementary DNA confers nitrosourea resistance to human hematopoietic progenitors." Clin Cancer Res 2(8): 1353-9.

Allay, J. A., D. A. Persons, J. Galipeau, J. M. Riberdy, R. A. Ashmun, R. L. Blakley, et al. (1998). "In vivo selection of retrovirally transduced hematopoietic stem cells." Nat Med 4(10): 1136-43.

156

Allay, J. A., H. T. Spencer, S. L. Wilkinson, J. A. Belt, R. L. Blakley and B. P. Sorrentino (1997). "Sensitization of hematopoietic stem and progenitor cells to trimetrexate using nucleoside transport inhibitors." Blood 90(9): 3546-54.

Amendola, M., M. A. Venneri, A. Biffi, E. Vigna and L. Naldini (2005). "Coordinate dual-gene transgenesis by lentiviral vectors carrying synthetic bidirectional promoters." Nat Biotechnol 23(1): 108-16 Epub 2004 Dec 26.

Amsellem, S., F. Pflumio, D. Bardinet, B. Izac, P. Charneau, P. H. Romeo, et al. (2003). "Ex vivo expansion of human hematopoietic stem cells by direct delivery of the HOXB4 homeoprotein." Nat Med 9(11): 1423-7 Epub 2003 Oct 26.

An, D. S., R. P. Wersto, B. A. Agricola, M. E. Metzger, S. Lu, R. G. Amado, et al. (2000). "Marking and gene expression by a lentivirus vector in transplanted human and nonhuman primate CD34(+) cells." J Virol 74(3): 1286-95.

Antonchuk, J., G. Sauvageau and R. K. Humphries (2002). "HOXB4-induced expansion of adult hematopoietic stem cells ex vivo." Cell 109(1): 39-45.

Artelt, P., C. Morelle, M. Ausmeier, M. Fitzek and H. Hauser (1988). "Vectors for efficient expression in mammalian fibroblastoid, myeloid and lymphoid cells via transfection or infection." Gene 68(2): 213-9.

Attal, J., M. C. Theron and L. M. Houdebine (1999). "The optimal use of IRES (internal ribosome entry site) in expression vectors." Genet Anal 15(3-5): 161-5.

Bae, Y., S. M. Kingsman and A. J. Kingsman (1997). "Functional dissection of the Moloney murine leukemia virus envelope protein gp70." J Virol 71(3): 2092-9.

Battini, J.-L. (1995). "Receptor-Binding Domain of Murine Leukemia Virus Envelope Glycoproteins." Journal of Virology 69(2): 713-719.

Baum, C., A. Richters and W. Ostertag (1999). "Retroviral vector-mediated gene expression in hematopoietic cells." Curr Opin Mol Ther 1(5): 605-12.

157 Beausejour, C. M., N. Eliopoulos, L. Momparler, N. L. Le and R. L. Momparler (2001). "Selection of drug-resistant transduced cells with cytosine nucleoside analogs using the human cytidine deaminase gene." Cancer Gene Ther 8(9): 669-76.

Bellantuono, I., L. S. Lashford, J. A. Rafferty and L. J. Fairbairn (2000). "The expression of full length Gp91-phox protein is associated with reduced amphotropic retroviral production." Haematologica 85(5): 451-7.

Beslu, N., J. Krosl, M. Laurin, N. Mayotte, K. R. Humphries and G. Sauvageau (2004). "Molecular interactions involved in HOXB4-induced activation of HSC self-renewal." Blood 104(8): 2307-14 Epub 2004 Jun 29.

Blau, C. A., K. R. Peterson, J. G. Drachman and D. M. Spencer (1997). "A proliferation switch for genetically modified cells." Proc Natl Acad Sci U S A 94(7): 3076-81.

Bowtell, D. D., S. Cory, G. R. Johnson and T. J. Gonda (1988). "Comparison of expression in hemopoietic cells by retroviral vectors carrying two genes." J Virol 62(7): 2464-73.

Brennand, J. and G. P. Margison (1986). "Reduction of the toxicity and mutagenicity of alkylating agents in mammalian cells harboring the Escherichia coli alkyltransferase gene." Proc Natl Acad Sci U S A 83(17): 6292-6.

Brent, T. P., P. J. Houghton and J. A. Houghton (1985). "O6-Alkylguanine-DNA alkyltransferase activity correlates with the therapeutic response of human rhabdomyosarcoma xenografts to 1-(2-chloroethyl)-3-(trans-4-methylcyclohexyl)-1- nitrosourea." Proc Natl Acad Sci U S A 82(9): 2985-9.

Bunting, K. D., S. Zhou, T. Lu and B. P. Sorrentino (2000). "Enforced P-glycoprotein pump function in murine bone marrow cells results in expansion of side population stem cells in vitro and repopulating cells in vivo." Blood 96(3): 902-9.

Bupp, K. and M. J. Roth (2002). "Altering retroviral tropism using a random-display envelope library." Mol Ther 5(3): 329-35.

158 Carneiro, F. A., M. L. Bianconi, G. Weissmuller, F. Stauffer and A. T. Da Poian (2002). "Membrane recognition by vesicular stomatitis virus involves enthalpy-driven protein- lipid interactions." J Virol 76(8): 3756-64.

Carpinteiro, A., S. Peinert, W. Ostertag, A. R. Zander, D. K. Hossfeld, K. Kuhlcke, et al. (2002). "Genetic protection of repopulating hematopoietic cells with an improved MDR1-retrovirus allows administration of intensified chemotherapy following stem cell transplantation in mice." Int J Cancer 98(5): 785-92.

Chandrashekran, A., M. Y. Gordon and C. Casimir (2004). "Targeted retroviral transduction of c-kit+ hematopoietic cells using novel ligand display technology." Blood 104(9): 2697-703 Epub 2004 Jul 15.

Cherry, S. R., D. Biniszkiewicz, L. van Parijs, D. Baltimore and R. Jaenisch (2000). "Retroviral expression in embryonic stem cells and hematopoietic stem cells." Mol Cell Biol 20(20): 7419-26.

Chinnasamy, N., J. A. Rafferty, I. Hickson, L. S. Lashford, S. J. Longhurst, N. Thatcher, et al. (1998). "Chemoprotective gene transfer II: multilineage in vivo protection of haemopoiesis against the effects of an antitumour agent by expression of a mutant human O6-alkylguanine-DNA alkyltransferase." Gene Ther 5(6): 842-7.

Choi, J. K., N. Hoang, A. M. Vilardi, P. Conrad, S. G. Emerson and A. M. Gewirtz (2001). "Hybrid HIV/MSCV LTR enhances transgene expression of lentiviral vectors in human CD34(+) hematopoietic cells." Stem Cells 19(3): 236-46.

Choi, U., S. S. DeRavin, K. Yamashita, N. Whiting-Theobald, G. F. Linton, N. A. Loktionova, et al. (2004). "Nuclear-localizing O6-benzylguanine-resistant GFP-MGMT fusion protein as a novel in vivo selection marker." Exp Hematol 32(8): 709-19.

Christians, F. C., B. J. Dawson, M. M. Coates and L. A. Loeb (1997). "Creation of human alkyltransferases resistant to O6-benzylguanine." Cancer Res 57(10): 2007-12.

Chu, T. H. and R. Dornburg (1995). "Retroviral vector particles displaying the antigen- binding site of an antibody enable cell-type-specific gene transfer." J Virol 69(4): 2659- 2663.

159 Citron, M., R. Decker, S. Chen, S. Schneider, M. Graver, L. Kleynerman, et al. (1991). "O6-methylguanine-DNA methyltransferase in human normal and tumor tissue from brain, lung, and ovary." Cancer Res 51(16): 4131-4.

Corey, C. A., A. D. DeSilva, C. A. Holland and D. A. Williams (1990). "Serial transplantation of methotrexate-resistant bone marrow: protection of murine recipients from drug toxicity by progeny of transduced stem cells." Blood 75(2): 337-43.

Cosset, F.-L. (1995). "Retroviral Retargeting by Envelopes Expressing an N-Terminal Binding Domain." Journal of Virology 69(10): 6314-6322.

Cowan, K. H., J. A. Moscow, H. Huang, J. A. Zujewski, J. O'Shaughnessy, B. Sorrentino, et al. (1999). "Paclitaxel chemotherapy after autologous stem-cell transplantation and engraftment of hematopoietic cells transduced with a retrovirus containing the multidrug resistance complementary DNA (MDR1) in metastatic breast cancer patients." Clin Cancer Res 5(7): 1619-28.

Creancier, L., D. Morello, P. Mercier and A. C. Prats (2000). "Fibroblast growth factor 2 internal ribosome entry site (IRES) activity ex vivo and in transgenic mice reveals a stringent tissue-specific regulation." J Cell Biol 150(1): 275-81.

Crone, T. M., K. Goodtzova, S. Edara and A. E. Pegg (1994). "Mutations in human O6- alkylguanine-DNA alkyltransferase imparting resistance to O6-benzylguanine." Cancer Res 54(23): 6221-7.

Dao, M. A., C. H. Hannum, D. B. Kohn and J. A. Nolta (1997). "FLT3 ligand preserves the ability of human CD34+ progenitors to sustain long-term hematopoiesis in immune- deficient mice after ex vivo retroviral-mediated transduction." Blood 89(2): 446-56.

Davey, R. A., Y. Zuo and J. M. Cunningham (1999). "Identification of a receptor-binding pocket on the envelope protein of friend murine leukemia virus." J Virol 73(5): 3758- 3763.

Davies, M. V. and R. J. Kaufman (1992). "The sequence context of the initiation codon in the encephalomyocarditis virus leader modulates efficiency of internal translation initiation." J Virol 66(4): 1924-32.

160 Davis, B. M., O. N. Koc and S. L. Gerson (2000). "Limiting numbers of G156A O(6)- methylguanine-DNA methyltransferase-transduced marrow progenitors repopulate nonmyeloablated mice after drug selection." Blood 95(10): 3078-84.

Davis, B. M., J. S. Reese, O. N. Koc, K. Lee, J. E. Schupp and S. L. Gerson (1997). "Selection for G156A O6-methylguanine DNA methyltransferase gene-transduced hematopoietic progenitors and protection from lethality in mice treated with O6- benzylguanine and 1,3-bis(2-chloroethyl)-1-nitrosourea." Cancer Res 57(22): 5093-9.

de Felipe, P. (2004). "Skipping the co-expression problem: the new 2A "CHYSEL" technology." Genet Vaccines Ther 2(1): 13.

de Felipe, P., V. Martin, M. L. Cortes, M. Ryan and M. Izquierdo (1999). "Use of the 2A sequence from foot-and-mouth disease virus in the generation of retroviral vectors for gene therapy." Gene Ther 6(2): 198-208.

Demaison, C., K. Parsley, G. Brouns, M. Scherr, K. Battmer, C. Kinnon, et al. (2002). "High-level transduction and gene expression in hematopoietic repopulating cells using a human immunodeficiency [correction of imunodeficiency] virus type 1-based lentiviral vector containing an internal spleen focus forming virus promoter." Hum Gene Ther 13(7): 803-13.

Dolan, M. E., R. C. Moschel and A. E. Pegg (1990). "Depletion of mammalian O6- alkylguanine-DNA alkyltransferase activity by O6-benzylguanine provides a means to evaluate the role of this protein in protection against carcinogenic and therapeutic alkylating agents." Proc Natl Acad Sci U S A 87(14): 5368-72.

Dolan, M. E., A. E. Pegg, N. K. Hora and L. C. Erickson (1988). "Effect of O6- methylguanine on DNA interstrand cross-link formation by chloroethylnitrosoureas and 2-chloroethyl(methylsulfonyl)methanesulfonate." Cancer Res 48(13): 3603-6.

Dolan, M. E., A. E. Pegg, R. C. Moschel and G. B. Grindey (1993). "Effect of O6- benzylguanine on the sensitivity of human colon tumor xenografts to 1,3-bis(2- chloroethyl)-1-nitrosourea (BCNU)." Biochem Pharmacol 46(2): 285-90.

Douglas, J. L., W. Y. Lin, M. L. Panis and G. Veres (2001). "Efficient human immunodeficiency virus-based vector transduction of unstimulated human mobilized

161 peripheral blood CD34+ cells in the SCID-hu Thy/Liv model of human T cell lymphopoiesis." Hum Gene Ther 12(4): 401-13.

Eliopoulos, N., V. Bovenzi, N. L. Le, L. F. Momparler, M. Greenbaum, S. Letourneau, et al. (1998). "Retroviral transfer and long-term expression of human cytidine deaminase cDNA in hematopoietic cells following transplantation in mice." Gene Ther 5(11): 1545- 51.

Emerman, M. and H. M. Temin (1984). "Genes with promoters in retrovirus vectors can be independently suppressed by an epigenetic mechanism." Cell 39(3 Pt 2): 449-67.

Emery, D. W., E. Yannaki, J. Tubb and G. Stamatoyannopoulos (2000). "A chromatin insulator protects retrovirus vectors from chromosomal position effects." Proc Natl Acad Sci U S A 97(16): 9150-5.

Etienne-Julan, M., P. Roux, P. Bourquard, S. Carillo, P. Jeanteur and M. Piechaczyk (1992). "Cell targeting by murine recombinant retroviruses." Bone Marrow Transplant 9: 139-142.

Etienne-Julan, M., P. Roux, S. Carillo, P. Jeanteur and M. Piechaczyk (1992). "The efficiency of cell targeting by recombinant retroviruses depends on the nature of the receptor and the composition of the artificial cell-virus linker." J Gen Virol 73(Pt 12): 3251-5.

Fairbairn, L. J., A. J. Watson, J. A. Rafferty, R. H. Elder and G. P. Margison (1995). "O6-benzylguanine increases the sensitivity of human primary bone marrow cells to the cytotoxic effects of temozolomide." Exp Hematol 23(2): 112-6.

Farson, D., R. Witt, R. McGuinness, T. Dull, M. Kelly, J. Song, et al. (2001). "A new- generation stable inducible packaging cell line for lentiviral vectors." Hum Gene Ther 12(8): 981-97.

Fass, D., R. A. Davey, C. A. Hamson, P. S. Kim, J. M. Cunningham and J. M. Berger (1997). "Structure of a murine leukemia virus receptor-binding glycoprotein at 2.0 angstrom resolution." Science 277(5332): 1662-6.

162 Follenzi, A., L. E. Ailles, S. Bakovic, M. Geuna and L. Naldini (2000). "Gene transfer by lentiviral vectors is limited by nuclear translocation and rescued by HIV-1 pol sequences." Nat Genet 25(2): 217-22.

Frimpong, K. and S. A. Spector (2000). "Cotransduction of nondividing cells using lentiviral vectors." Gene Ther 7(18): 1562-9.

Furler, S., J. C. Paterna, M. Weibel and H. Bueler (2001). "Recombinant AAV vectors containing the foot and mouth disease virus 2A sequence confer efficient bicistronic gene expression in cultured cells and rat substantia nigra neurons." Gene Ther 8(11): 864-73.

Galski, H., M. Sullivan, M. C. Willingham, K. V. Chin, M. M. Gottesman, I. Pastan, et al. (1989). "Expression of a human multidrug resistance cDNA (MDR1) in the bone marrow of transgenic mice: resistance to daunomycin-induced leukopenia." Mol Cell Biol 9(10): 4357-63.

Gerson, S. L., K. Miller and N. A. Berger (1985). "O6 alkylguanine-DNA alkyltransferase activity in human myeloid cells." J Clin Invest 76(6): 2106-14.

Gerson, S. L., J. E. Trey, K. Miller and E. Benjamin (1987). "Repair of O6-alkylguanine during DNA synthesis in murine bone marrow hematopoietic precursors." Cancer Res 47(1): 89-95.

Giampaolo, A., P. Sterpetti, D. Bulgarini, P. Samoggia, E. Pelosi, M. Valtieri, et al. (1994). "Key functional role and lineage-specific expression of selected HOXB genes in purified hematopoietic progenitor differentiation." Blood 84(11): 3637-47.

Gollan, T. J. and M. R. Green (2002). "Redirecting retroviral tropism by insertion of short, nondisruptive peptide ligands into envelope." J Virol 76(7): 3558-63.

Gonzaga, P. E. and T. P. Brent (1989). "Affinity purification and characterization of human O6-alkylguanine-DNA alkyltransferase complexed with BCNU-treated, synthetic oligonucleotide." Nucleic Acids Res 17(16): 6581-90.

Goodell, M. A., K. Brose, G. Paradis, A. S. Conner and R. C. Mulligan (1996). "Isolation and functional properties of murine hematopoietic stem cells that are replicating in vivo." J Exp Med 183(4): 1797-806.

163

Gothot, A., J. C. van der Loo, D. W. Clapp and E. F. Srour (1998). "Cell cycle-related changes in repopulating capacity of human mobilized peripheral blood CD34(+) cells in non-obese diabetic/severe combined immune-deficient mice." Blood 92(8): 2641-9.

Grier, D. G., A. Thompson, A. Kwasniewska, G. J. McGonigle, H. L. Halliday and T. R. Lappin (2005). "The pathophysiology of HOX genes and their role in cancer." J Pathol 205(2): 154-71.

Hacein-Bey-Abina, S., C. von Kalle, M. Schmidt, F. Le Deist, N. Wulffraat, E. McIntyre, et al. (2003). "A serious adverse event after successful gene therapy for X-linked severe combined immunodeficiency." N Engl J Med 348(3): 255-6.

Hafkemeyer, P., T. Licht, I. Pastan and M. M. Gottesman (2000). "Chemoprotection of hematopoietic cells by a mutant P-glycoprotein resistant to a potent chemosensitizer of multidrug-resistant cancers." Hum Gene Ther 11(4): 555-65.

Halene, S., L. Wang, R. M. Cooper, D. C. Bockstoce, P. B. Robbins and D. B. Kohn (1999). "Improved expression in hematopoietic and lymphoid cells in mice after transplantation of bone marrow transduced with a modified retroviral vector." Blood 94(10): 3349-57.

Han, J. Y., P. M. Cannon, K. M. Lai, Y. Zhao, M. V. Eiden and W. F. Anderson (1997). "Identification of envelope protein residues required for the expanded host range of 10A1 murine leukemia virus." J Virol 71(11): 8103-8.

Han, J. Y., Y. Zhao, W. F. Anderson and P. M. Cannon (1998). "Role of variable regions A and B in receptor binding domain of amphotropic murine leukemia virus envelope protein." J Virol 72(11): 9101-8.

Han, X., N. Kasahara and Y. W. Kan (1995). "Ligand-directed retroviral targeting of human breast cancer cells." Proc Natl Acad Sci U S A 92(21): 9747-9751.

Hanania, E. G. and A. B. Deisseroth (1994). "Serial transplantation shows that early hematopoietic precursor cells are transduced by MDR-1 retroviral vector in a mouse gene therapy model." Cancer Gene Ther 1(1): 21-5.

164

Hanania, E. G. and A. B. Deisseroth (1997). "Simultaneous genetic chemoprotection of normal marrow cells and genetic chemosensitization of breast cancer cells in a mouse cancer gene therapy model." Clin Cancer Res 3(2): 281-6.

Hanania, E. G., R. E. Giles, J. Kavanagh, S. Q. Fu, D. Ellerson, Z. Zu, et al. (1996). "Results of MDR-1 vector modification trial indicate that granulocyte/macrophage colony-forming unit cells do not contribute to posttransplant hematopoietic recovery following intensive systemic therapy." Proc Natl Acad Sci U S A 93(26): 15346-51.

Hanawa, H., P. F. Kelly, A. C. Nathwani, D. A. Persons, J. A. Vandergriff, P. Hargrove, et al. (2002). "Comparison of various envelope proteins for their ability to pseudotype lentiviral vectors and transduce primitive hematopoietic cells from human blood." Mol Ther 5(3): 242-51.

Harris, L. C., U. K. Marathi, C. C. Edwards, P. J. Houghton, D. K. Srivastava, E. F. Vanin, et al. (1995). "Retroviral transfer of a bacterial alkyltransferase gene into murine bone marrow protects against chloroethylnitrosourea cytotoxicity." Clin Cancer Res 1(11): 1359-68.

Hawley, R. G., A. Z. Fong, B. F. Burns and T. S. Hawley (1992). "Transplantable myeloproliferative disease induced in mice by an interleukin 6 retrovirus." J Exp Med 176(4): 1149-63.

Hennecke, M., M. Kwissa, K. Metzger, A. Oumard, A. Kroger, R. Schirmbeck, et al. (2001). "Composition and arrangement of genes define the strength of IRES-driven translation in bicistronic mRNAs." Nucleic Acids Res 29(16): 3327-34.

Hesdorffer, C., J. Ayello, M. Ward, A. Kaubisch, L. Vahdat, C. Balmaceda, et al. (1998). "Phase I trial of retroviral-mediated transfer of the human MDR1 gene as marrow chemoprotection in patients undergoing high-dose chemotherapy and autologous stem- cell transplantation." J Clin Oncol 16(1): 165-72.

Hickson, I., L. J. Fairbairn, N. Chinnasamy, L. S. Lashford, N. Thatcher, G. P. Margison, et al. (1998). "Chemoprotective gene transfer I: transduction of human haemopoietic progenitors with O6-benzylguanine-resistant O6-alkylguanine-DNA alkyltransferase attenuates the toxic effects of O6-alkylating agents in vitro." Gene Ther 5(6): 835-41.

165 Hildinger, M., B. Fehse, S. Hegewisch-Becker, J. John, J. R. Rafferty, W. Ostertag, et al. (1998). "Dominant selection of hematopoietic progenitor cells with retroviral MDR1 co- expression vectors." Hum Gene Ther 9(1): 33-42.

Horton, R. M., H. D. Hunt, S. N. Ho, J. K. Pullen and L. R. Pease (1989). "Engineering hybrid genes without the use of restriction enzymes: gene splicing by overlap extension." Gene 77(1): 61-8.

Inoue, T., H. Yamaza, Y. Sakai, S. Mizuno, M. Ohno, N. Hamasaki, et al. (1999). "Position-independent human beta-globin gene expression mediated by a recombinant adeno-associated virus vector carrying the chicken beta- globin insulator." J Hum Genet 44(3): 152-62.

Ishizaki, K., T. Tsujimura, H. Yawata, C. Fujio, Y. Nakabeppu, M. Sekiguchi, et al. (1986). "Transfer of the E. coli O6-methylguanine methyltransferase gene into repair- deficient human cells and restoration of cellular resistance to N-methyl-N'-nitro-N- nitrosoguanidine." Mutat Res 166(2): 135-41.

Jelinek, J., K. Kleibl, T. M. Dexter and G. P. Margison (1988). "Transfection of murine multi-potent haemopoietic stem cells with an E. coli DNA alkyltransferase gene confers resistance to the toxic effects of alkylating agents." Carcinogenesis 9(1): 81-7.

Jelinek, J., J. A. Rafferty, R. Cmejla, M. Hildinger, D. Chinnasamy, L. S. Lashford, et al. (1999). "A novel dual function retrovirus expressing multidrug resistance 1 and O6- alkylguanine-DNA-alkyltransferase for engineering resistance of haemopoietic progenitor cells to multiple chemotherapeutic agents." Gene Ther 6(8): 1489-93.

Jia, L., L. Cheng and J. Raper (2005). "Slit/Robo signaling is necessary to confine early neural crest cells to the ventral migratory pathway in the trunk." Dev Biol 282(2): 411- 21.

Jiang, A., T. H. Chu, F. Nocken, K. Cichutek and R. Dornburg (1998). "Cell-type- specific gene transfer into human cells with retroviral vectors that display single-chain antibodies." J Virol 72(12): 10148-56.

Jin, L., H. Asano and C. A. Blau (1998). "Stimulating cell proliferation through the pharmacologic activation of c-kit." Blood 91(3): 890-7.

166

Jin, L., N. Siritanaratkul, D. W. Emery, R. E. Richard, K. Kaushansky, T. Papayannopoulou, et al. (1998). "Targeted expansion of genetically modified bone marrow cells." Proc Natl Acad Sci U S A 95(14): 8093-7.

Juliano, R. L. and V. Ling (1976). "A surface glycoprotein modulating drug permeability in Chinese hamster ovary cell mutants." Biochim Biophys Acta 455(1): 152-62.

Kaiser, J. (2003). "Gene therapy. Seeking the cause of induced leukemias in X-SCID trial." Science 299(5606): 495.

Kane, S. E., L. Matsumoto, M. Z. Metz, C. A. Donohue, R. G. Lingeman, S. W. Lin, et al. (2001). "MDR1 bicistronic vectors: analysis of selection stringency, amplified gene expression, and vector stability in cell lines." Biochem Pharmacol 62(6): 693-704.

Karran, P. and M. Bignami (1994). "DNA damage tolerance, mismatch repair and genome instability." Bioessays 16(11): 833-9.

Kartner, N., D. Evernden-Porelle, G. Bradley and V. Ling (1985). "Detection of P- glycoprotein in multidrug-resistant cell lines by monoclonal antibodies." Nature 316(6031): 820-3.

Kasahara, N. (1994). "Tissue-Specific Targeting of Retroviral Vectors Through Ligand- Receptor Interactions." Science 266: 1373-1376.

Kavanaugh, M. P., D. G. Miller, W. Zhang, W. Law, S. L. Kozak, D. Kabat, et al. (1994). "Cell-surface receptors for gibbon ape leukemia virus and amphotropic murine retrovirus are inducible sodium-dependent phosphate symporters." Proc Natl Acad Sci U S A 91(15): 7071-5.

Kelly, P. F., J. Carrington, A. Nathwani and E. F. Vanin (2001). "RD114-pseudotyped oncoretroviral vectors. Biological and physical properties." Ann N Y Acad Sci 938: 262- 76; discussion 276-7.

Ketteler, R., S. Glaser, O. Sandra, U. M. Martens and U. Klingmuller (2002). "Enhanced transgene expression in primitive hematopoietic progenitor cells and embryonic stem

167 cells efficiently transduced by optimized retroviral hybrid vectors." Gene Ther 9(8): 477- 87.

Kim, J. W., E. I. Closs, L. M. Albritton and J. M. Cunningham (1991). "Transport of cationic amino acids by the mouse ecotropic retrovirus receptor." Nature 352(6337): 725- 8.

Kim, M., H. Turnquist, J. Jackson, M. Sgagias, Y. Yan, M. Gong, et al. (2002). "The multidrug resistance transporter ABCG2 (breast cancer resistance protein 1) effluxes Hoechst 33342 and is overexpressed in hematopoietic stem cells." Clin Cancer Res 8(1): 22-8.

Kittler, E. L., S. O. Peters, R. B. Crittenden, M. E. Debatis, H. S. Ramshaw, F. M. Stewart, et al. (1997). "Cytokine-facilitated transduction leads to low-level engraftment in nonablated hosts." Blood 90(2): 865-72.

Koc, O. N., J. S. Reese, B. M. Davis, L. Liu, K. J. Majczenko and S. L. Gerson (1999). "DeltaMGMT-transduced bone marrow infusion increases tolerance to O6-benzylguanine and 1,3-bis(2-chloroethyl)-1-nitrosourea and allows intensive therapy of 1,3-bis(2- chloroethyl)-1-nitrosourea-resistant human colon cancer xenografts." Hum Gene Ther 10(6): 1021-30.

Konishi, H. (1998). "Targeting Strategy for Gene Delivery to Carcinoembryonic Antigen- Producing Cancer Cells by Retrovirus Displaying a Sigle-Chain Fragment Antibody." Human Gene Therapy 9: 235-248.

Kreklau, E. L., K. E. Pollok, B. J. Bailey, N. Liu, J. R. Hartwell, D. A. Williams, et al. (2003). "Hematopoietic expression of O(6)-methylguanine DNA methyltransferase- P140K allows intensive treatment of human glioma xenografts with combination O(6)- benzylguanine and 1,3-bis-(2-chloroethyl)-1-nitrosourea." Mol Cancer Ther 2(12): 1321- 9.

Krosl, J., P. Austin, N. Beslu, E. Kroon, R. K. Humphries and G. Sauvageau (2003). "In vitro expansion of hematopoietic stem cells by recombinant TAT-HOXB4 protein." Nat Med 9(11): 1428-32 Epub 2003 Oct 26.

Kumar, M., B. Keller, N. Makalou and R. E. Sutton (2001). "Systematic determination of the packaging limit of lentiviral vectors." Hum Gene Ther 12(15): 1893-905.

168

Kustikova, O. S., A. Wahlers, K. Kuhlcke, B. Stahle, A. R. Zander, C. Baum, et al. (2003). "Dose finding with retroviral vectors: correlation of retroviral vector copy numbers in single cells with gene transfer efficiency in a cell population." Blood 102(12): 3934-7 Epub 2003 Jul 24.

Laufs, S., B. Gentner, K. Z. Nagy, A. Jauch, A. Benner, S. Naundorf, et al. (2003). "Retroviral vector integration occurs in preferred genomic targets of human bone marrow-repopulating cells." Blood 101(6): 2191-8.

Leboulch, P., G. M. Huang, R. K. Humphries, Y. H. Oh, C. J. Eaves, D. Y. Tuan, et al. (1994). "Mutagenesis of retroviral vectors transducing human beta-globin gene and beta- globin locus control region derivatives results in stable transmission of an active transcriptional structure." Embo J 13(13): 3065-76.

Lengler, J., H. Holzmuller, B. Salmons, W. H. Gunzburg and M. Renner (2005). "FMDV-2A sequence and protein arrangement contribute to functionality of CYP2B1- reporter fusion protein." Anal Biochem 12: 12.

Lewis, W. S., V. Cody, N. Galitsky, J. R. Luft, W. Pangborn, S. K. Chunduru, et al. (1995). "Methotrexate-resistant variants of human dihydrofolate reductase with substitutions of leucine 22. Kinetics, crystallography, and potential as selectable markers." J Biol Chem 270(10): 5057-64.

Lotti, F., E. Menguzzato, C. Rossi, L. Naldini, L. Ailles, F. Mavilio, et al. (2002). "Transcriptional targeting of lentiviral vectors by long terminal repeat enhancer replacement." J Virol 76(8): 3996-4007.

Luens, K. M., M. A. Travis, B. P. Chen, B. L. Hill, R. Scollay and L. J. Murray (1998). "Thrombopoietin, kit ligand, and flk2/flt3 ligand together induce increased numbers of primitive hematopoietic progenitors from human CD34+Thy-1+Lin- cells with preserved ability to engraft SCID-hu bone." Blood 91(4): 1206-15.

MacKrell, A. J., N. W. Soong, C. M. Curtis and W. F. Anderson (1996). "Identification of a subdomain in the Moloney murine leukemia virus envelope protein involved in receptor binding." J Virol 70(3): 1768-74.

169 Manilla, P., T. Rebello, C. Afable, X. Lu, V. Slepushkin, L. M. Humeau, et al. (2005). "Regulatory considerations for novel gene therapy products: a review of the process leading to the first clinical lentiviral vector." Hum Gene Ther 16(1): 17-25.

Margison, G. P., D. P. Cooper and J. Brennand (1985). "Cloning of the E. coli O6- methylguanine and methylphosphotriester methyltransferase gene using a functional DNA repair assay." Nucleic Acids Res 13(6): 1939-52.

Marin, M. (1996). "Targeted Infection of Human Cells via Major Histocompatibility Complex Class I Molecules by Moloney Murine Leukemia Virus-Derived Viruses Displaying Single-Chain Antibody Fragment-Envelope Fusion Proteins." Journal of Virology 70(5): 2957-2962.

Mickisch, G. H., I. Aksentijevich, P. V. Schoenlein, L. J. Goldstein, H. Galski, C. Stahle, et al. (1992). "Transplantation of bone marrow cells from transgenic mice expressing the human MDR1 gene results in long-term protection against the myelosuppressive effect of chemotherapy in mice." Blood 79(4): 1087-93.

Mickisch, G. H., T. Licht, G. T. Merlino, M. M. Gottesman and I. Pastan (1991). "Chemotherapy and chemosensitization of transgenic mice which express the human multidrug resistance gene in bone marrow: efficacy, potency, and toxicity." Cancer Res 51(19): 5417-24.

Miller, D. (1996). "Cell-surface receptors for retroviruses and implications for gene transfer." Proc. Natl. Acad. Sci. USA 93: 11407-11413.

Miller, D. G., R. H. Edwards and A. D. Miller (1994). "Cloning of the cellular receptor for amphotropic murine retroviruses reveals homology to that for gibbon ape leukemia virus." Proc Natl Acad Sci U S A 91(1): 78-82.

Milsom, M. D., L. B. Woolford, G. P. Margison, R. K. Humphries and L. J. Fairbairn (2004). "Enhanced in vivo selection of bone marrow cells by retroviral-mediated coexpression of mutant O6-methylguanine-DNA-methyltransferase and HOXB4." Mol Ther 10(5): 862-73.

Miyoshi, H., U. Blomer, M. Takahashi, F. H. Gage and I. M. Verma (1998). "Development of a self-inactivating lentivirus vector." J Virol 72(10): 8150-7.

170

Modlich, U., O. S. Kustikova, M. Schmidt, C. Rudolph, J. Meyer, Z. Li, et al. (2005). "Leukemias following retroviral transfer of multidrug resistance 1 (MDR1) are driven by combinatorial insertional mutagenesis." Blood 15: 15.

Moreau-Gaudry, F., P. Xia, G. Jiang, N. P. Perelman, G. Bauer, J. Ellis, et al. (2001). "High-level erythroid-specific gene expression in primary human and murine hematopoietic cells with self-inactivating lentiviral vectors." Blood 98(9): 2664-72.

Moretti, P., P. Simmons, P. Thomas, D. Haylock, P. Rathjen, M. Vadas, et al. (1994). "Identification of homeobox genes expressed in human haemopoietic progenitor cells." Gene 144(2): 213-9.

Moritz, T., P. Dutt, X. Xiao, D. Carstanjen, T. Vik, H. Hanenberg, et al. (1996). "Fibronectin improves transduction of reconstituting hematopoietic stem cells by retroviral vectors: evidence of direct viral binding to chymotryptic carboxy-terminal fragments." Blood 88(3): 855-62.

Moritz, T., W. Mackay, B. J. Glassner, D. A. Williams and L. Samson (1995). "Retrovirus-mediated expression of a DNA repair protein in bone marrow protects hematopoietic cells from nitrosourea-induced toxicity in vitro and in vivo." Cancer Res 55(12): 2608-14.

Moriwaki, S., T. Yagi, C. Nishigori, S. Imamura and H. Takebe (1991). "Analysis of N- methyl-N-nitrosourea-induced mutations in a shuttle vector plasmid propagated in mouse O6-methylguanine-DNA methyltransferase-deficient cells in comparison with proficient cells." Cancer Res 51(23 Pt 1): 6219-23.

Moscow, J. A., H. Huang, C. Carter, K. Hines, J. Zujewski, G. Cusack, et al. (1999). "Engraftment of MDR1 and NeoR gene-transduced hematopoietic cells after breast cancer chemotherapy." Blood 94(1): 52-61.

Murray, L., M. Travis, K. Luens-Abitorabi, K. Olsson, I. Plavec, S. Forestell, et al. (2000). "Addition of the human interferon beta scaffold attachment region to retroviral vector backbones increases the level of in vivo transgene expression among progeny of engrafted human hematopoietic stem cells." Hum Gene Ther 11(14): 2039-50.

171 Naldini, L., U. Blomer, P. Gallay, D. Ory, R. Mulligan, F. H. Gage, et al. (1996). "In vivo gene delivery and stable transduction of nondividing cells by a lentiviral vector." Science 272(5259): 263-7.

Neff, T., B. C. Beard, L. J. Peterson, P. Anandakumar, J. Thompson and H. P. Kiem (2005). "Polyclonal chemoprotection against temozolomide in a large-animal model of drug resistance gene therapy." Blood 105(3): 997-1002 Epub 2004 Oct 19.

Neff, T., P. A. Horn, L. J. Peterson, B. M. Thomasson, J. Thompson, D. A. Williams, et al. (2003). "Methylguanine methyltransferase-mediated in vivo selection and chemoprotection of allogeneic stem cells in a large-animal model." J Clin Invest 112(10): 1581-8.

Nguyen, T. H., J. C. Pages, D. Farge, P. Briand and A. Weber (1998). "Amphotropic retroviral vectors displaying hepatocyte growth factor- envelope fusion proteins improve transduction efficiency of primary hepatocytes." Hum Gene Ther 9(17): 2469-2479.

Nowakowski, G. S., M. S. Dooner, H. M. Valinski, A. M. Mihaliak, P. J. Quesenberry and P. S. Becker (2004). "A specific heptapeptide from a phage display peptide library homes to bone marrow and binds to primitive hematopoietic stem cells." Stem Cells 22(6): 1030-8.

Pacchia, A. L., M. E. Adelson, M. Kaul, Y. Ron and J. P. Dougherty (2001). "An inducible packaging cell system for safe, efficient lentiviral vector production in the absence of HIV-1 accessory proteins." Virology 282(1): 77-86.

Pasqualini, R. and E. Ruoslahti (1996). "Organ targeting in vivo using phage display peptide libraries." Nature 380(6572): 364-6.

Pawliuk, R., K. A. Westerman, M. E. Fabry, E. Payen, R. Tighe, E. E. Bouhassira, et al. (2001). "Correction of sickle cell disease in transgenic mouse models by gene therapy." Science 294(5550): 2368-71.

Pegg, A. E., K. Swenn, M. Y. Chae, M. E. Dolan and R. C. Moschel (1995). "Increased killing of prostate, breast, colon, and lung tumor cells by the combination of inactivators of O6-alkylguanine-DNA alkyltransferase and N,N'-bis(2-chloroethyl)-N-nitrosourea." Biochem Pharmacol 50(8): 1141-8.

172

Pegg, A. E., L. Wiest, R. S. Foote, S. Mitra and W. Perry (1983). "Purification and properties of O6-methylguanine-DNA transmethylase from rat liver." J Biol Chem 258(4): 2327-33.

Persons, D. A., J. A. Allay, A. Bonifacino, T. Lu, B. Agricola, M. E. Metzger, et al. (2004). "Transient in vivo selection of transduced peripheral blood cells using antifolate drug selection in rhesus macaques that received transplants with hematopoietic stem cells expressing dihydrofolate reductase vectors." Blood 103(3): 796-803 Epub 2003 Aug 14.

Podda, S., M. Ward, A. Himelstein, C. Richardson, E. de la Flor-Weiss, L. Smith, et al. (1992). "Transfer and expression of the human multiple drug resistance gene into live mice." Proc Natl Acad Sci U S A 89(20): 9676-80.

Pollok, K. E., J. R. Hartwell, A. Braber, R. J. Cooper, M. Jansen, S. Ragg, et al. (2003). "In vivo selection of human hematopoietic cells in a xenograft model using combined pharmacologic and genetic manipulations." Hum Gene Ther 14(18): 1703-14.

Popa, I., M. E. Harris, J. E. Donello and T. J. Hope (2002). "CRM1-dependent function of a cis-acting RNA export element." Mol Cell Biol 22(7): 2057-67.

Prasad, A., A. Z. Fernandis, Y. Rao and R. K. Ganju (2004). "Slit protein-mediated inhibition of CXCR4-induced chemotactic and chemoinvasive signaling pathways in breast cancer cells." J Biol Chem 279(10): 9115-24 Epub 2003 Nov 26.

Ragg, S., M. Xu-Welliver, J. Bailey, M. D'Souza, R. Cooper, S. Chandra, et al. (2000). "Direct reversal of DNA damage by mutant methyltransferase protein protects mice against dose-intensified chemotherapy and leads to in vivo selection of hematopoietic stem cells." Cancer Res 60(18): 5187-95.

Randall, T. D. and I. L. Weissman (1997). "Phenotypic and functional changes induced at the clonal level in hematopoietic stem cells after 5-fluorouracil treatment." Blood 89(10): 3596-606.

Reese, J. S., B. M. Davis, L. Liu and S. L. Gerson (1999). "Simultaneous protection of G156A methylguanine DNA methyltransferase gene-transduced hematopoietic progenitors and sensitization of tumor cells using O6-benzylguanine and temozolomide." Clin Cancer Res 5(1): 163-9.

173

Reese, J. S., O. N. Koc, K. M. Lee, L. Liu, J. A. Allay, W. P. Phillips, Jr., et al. (1996). "Retroviral transduction of a mutant methylguanine DNA methyltransferase gene into human CD34 cells confers resistance to O6-benzylguanine plus 1,3-bis(2-chloroethyl)-1- nitrosourea." Proc Natl Acad Sci U S A 93(24): 14088-93.

Reese, J. S., K. T. Lingas, O. N. Koc and S. L. Gerson (2004). "Preliminary results of a phase I trial using retroviral gene transfer of G156A MGMT to protect hematopoiesis during BG and BCNU therapy of advanced malignancies." Mol Ther 9: S385.

Reese, J. S., L. Liu and S. L. Gerson (2003). "Repopulating defect of mismatch repair- deficient hematopoietic stem cells." Blood 102(5): 1626-33 Epub 2003 May 1.

Rein, A. (1990). "Sequence Analysis of Amphotropic and 10A1 Murine Leukemia Viruses: Close Relationship to Mink Cell Focus-Forming Viruses." Journal of Virology 64(2): 757-766.

Rein, A. and A. Schultz (1984). "Different recombinant murine leukemia viruses use different cell surface receptors." Virology 136(1): 144-52.

Reiser, J., G. Harmison, S. Kluepfel-Stahl, R. O. Brady, S. Karlsson and M. Schubert (1996). "Transduction of nondividing cells using pseudotyped defective high-titer HIV type 1 particles." Proc Natl Acad Sci U S A 93(26): 15266-71.

Reiser, J., Z. Lai, X. Y. Zhang and R. O. Brady (2000). "Development of multigene and regulated lentivirus vectors." J Virol 74(22): 10589-99.

Relander, T., M. Johansson, K. Olsson, Y. Ikeda, Y. Takeuchi, M. Collins, et al. (2005). "Gene Transfer to Repopulating Human CD34(+) Cells Using Amphotropic-, GALV-, or RD114-Pseudotyped HIV-1-Based Vectors from Stable Producer Cells." Mol Ther 11(3): 452-9.

Richard, E., E. Robert, M. Cario-Andre, C. Ged, F. Geronimi, S. L. Gerson, et al. (2004). "Hematopoietic stem cell gene therapy of murine protoporphyria by methylguanine- DNA-methyltransferase-mediated in vivo drug selection." Gene Ther 29: 29.

174 Richard, R. E., B. Wood, H. Zeng, L. Jin, T. Papayannopoulou and C. A. Blau (2000). "Expansion of genetically modified primary human hemopoietic cells using chemical inducers of dimerization." Blood 95(2): 430-6.

Rivella, S., J. A. Callegari, C. May, C. W. Tan and M. Sadelain (2000). "The cHS4 insulator increases the probability of retroviral expression at random chromosomal integration sites." J Virol 74(10): 4679-87.

Robbins, P. B., D. C. Skelton, X. J. Yu, S. Halene, E. H. Leonard and D. B. Kohn (1998). "Consistent, persistent expression from modified retroviral vectors in murine hematopoietic stem cells." Proc Natl Acad Sci U S A 95(17): 10182-7.

Roth, J. and S. Gerson (2002). "Screening mutant ecotropic MuLV envelope libraries for altered host range." Mol Ther 5(5): S177.

Roux, P., P. Jeanteur and M. Piechaczyk (1989). "A versatile and potentially general approach to the targeting of specific cell types by retroviruses: application to the infection of human cells by means of major histocompatibility complex class I and class II antigens by mouse ecotropic murine leukemia virus-derived viruses." Proc Natl Acad Sci U S A 86(23): 9079-83.

Russell, S. J. (1993). "Retroviral vectors displaying functional antibody fragments." Nucleic Acids Research 21(5): 1081-1085.

Ryan, M. D. and J. Drew (1994). "Foot-and-mouth disease virus 2A oligopeptide mediated cleavage of an artificial polyprotein." Embo J 13(4): 928-33.

Salmon, P., V. Kindler, O. Ducrey, B. Chapuis, R. H. Zubler and D. Trono (2000). "High-level transgene expression in human hematopoietic progenitors and differentiated blood lineages after transduction with improved lentiviral vectors." Blood 96(10): 3392- 8.

Samson, L., B. Derfler and E. A. Waldstein (1986). "Suppression of human DNA alkylation-repair defects by Escherichia coli DNA-repair genes." Proc Natl Acad Sci U S A 83(15): 5607-10.

175 Sauvageau, G., N. N. Iscove and R. K. Humphries (2004). "In vitro and in vivo expansion of hematopoietic stem cells." Oncogene 23(43): 7223-32.

Sauvageau, G., P. M. Lansdorp, C. J. Eaves, D. E. Hogge, W. H. Dragowska, D. S. Reid, et al. (1994). "Differential expression of homeobox genes in functionally distinct CD34+ subpopulations of human bone marrow cells." Proc Natl Acad Sci U S A 91(25): 12223- 7.

Sauvageau, G., U. Thorsteinsdottir, C. J. Eaves, H. J. Lawrence, C. Largman, P. M. Lansdorp, et al. (1995). "Overexpression of HOXB4 in hematopoietic cells causes the selective expansion of more primitive populations in vitro and in vivo." Genes Dev 9(14): 1753-65.

Scharenberg, C. W., M. A. Harkey and B. Torok-Storb (2002). "The ABCG2 transporter is an efficient Hoechst 33342 efflux pump and is preferentially expressed by immature human hematopoietic progenitors." Blood 99(2): 507-12.

Schiedlmeier, B., H. Klump, E. Will, G. Arman-Kalcek, Z. Li, Z. Wang, et al. (2003). "High-level ectopic HOXB4 expression confers a profound in vivo competitive growth advantage on human cord blood CD34+ cells, but impairs lymphomyeloid differentiation." Blood 101(5): 1759-68 Epub 2002 Oct 24.

Schiedlmeier, B., A. J. Schilz, K. Kuhlcke, S. Laufs, C. Baum, W. J. Zeller, et al. (2002). "Multidrug resistance 1 gene transfer can confer chemoprotection to human peripheral blood progenitor cells engrafted in immunodeficient mice." Hum Gene Ther 13(2): 233- 42.

Schmidt, M., H. Glimm, M. Wissler, G. Hoffmann, K. Olsson, S. Sellers, et al. (2003). "Efficient characterization of retro-, lenti-, and foamyvector-transduced cell populations by high-accuracy insertion site sequencing." Ann N Y Acad Sci 996: 112-21.

Schold, S. C., Jr., T. P. Brent, E. von Hofe, H. S. Friedman, S. Mitra, D. D. Bigner, et al. (1989). "O6-alkylguanine-DNA alkyltransferase and sensitivity to procarbazine in human brain-tumor xenografts." J Neurosurg 70(4): 573-7.

Sellers, S. E., J. F. Tisdale, B. A. Agricola, M. E. Metzger, R. E. Donahue, C. E. Dunbar, et al. (2001). "The effect of multidrug-resistance 1 gene versus neo transduction on ex

176 vivo and in vivo expansion of rhesus macaque hematopoietic repopulating cells." Blood 97(6): 1888-91.

Sirven, A., F. Pflumio, V. Zennou, M. Titeux, W. Vainchenker, L. Coulombel, et al. (2000). "The human immunodeficiency virus type-1 central DNA flap is a crucial determinant for lentiviral vector nuclear import and gene transduction of human hematopoietic stem cells." Blood 96(13): 4103-10.

Somia, N. V., M. Zoppe and I. M. Verma (1995). "Generation of targeted retroviral vectors by using single-chain variable fragment: an approach to in vivo gene delivery." Proc Natl Acad Sci U S A 92(16): 7570-7574.

Sorrentino, B. P., S. J. Brandt, D. Bodine, M. Gottesman, I. Pastan, A. Cline, et al. (1992). "Selection of drug-resistant bone marrow cells in vivo after retroviral transfer of human MDR1." Science 257(5066): 99-103.

Srivenugopal, K. S., X. H. Yuan, H. S. Friedman and F. Ali-Osman (1996). "Ubiquitination-dependent proteolysis of O6-methylguanine-DNA methyltransferase in human and murine tumor cells following inactivation with O6-benzylguanine or 1,3- bis(2-chloroethyl)-1-nitrosourea." Biochemistry 35(4): 1328-34.

Sutton, R. E., M. J. Reitsma, N. Uchida and P. O. Brown (1999). "Transduction of human progenitor hematopoietic stem cells by human immunodeficiency virus type 1-based vectors is cell cycle dependent." J Virol 73(5): 3649-60.

Szilvassy, S. J., T. E. Meyerrose, P. L. Ragland and B. Grimes (2001). "Differential homing and engraftment properties of hematopoietic progenitor cells from murine bone marrow, mobilized peripheral blood, and fetal liver." Blood 98(7): 2108-15.

Thomsen, S., B. Vogt, D. von Laer, C. Heberlein, A. Rein, W. Ostertag, et al. (1998). "Lack of functional Pit-1 and Pit-2 expression on hematopoietic stem cell lines." Acta Haematol 99(3): 148-55.

Thorsteinsdottir, U., G. Sauvageau and R. K. Humphries (1999). "Enhanced in vivo regenerative potential of HOXB4-transduced hematopoietic stem cells with regulation of their pool size." Blood 94(8): 2605-12.

177 Tucker, K. A., M. B. Lilly, L. Heck, Jr. and T. A. Rado (1987). "Characterization of a new human diploid myeloid leukemia cell line (PLB- 985) with granulocytic and monocytic differentiating capacity." Blood 70(2): 372-378.

Tumas, D. B., G. J. Spangrude, D. M. Brooks, C. D. Williams and B. Chesebro (1996). "High-frequency cell surface expression of a foreign protein in murine hematopoietic stem cells using a new retroviral vector." Blood 87(2): 509-17.

Uchida, N., F. Y. Leung and C. J. Eaves (2002). "Liver and marrow of adult mdr-1a/1b(- /-) mice show normal generation, function, and multi-tissue trafficking of primitive hematopoietic cells." Exp Hematol 30(8): 862-9.

Ueda, K., Y. Hanazono, H. Shibata, N. Ageyama, Y. Ueda, S. Ogata, et al. (2004). "High-level in vivo gene marking after gene-modified autologous hematopoietic stem cell transplantation without marrow conditioning in nonhuman primates." Mol Ther 10(3): 469-77.

Ujhelly, O., C. Ozvegy, G. Varady, J. Cervenak, L. Homolya, M. Grez, et al. (2003). "Application of a human multidrug transporter (ABCG2) variant as selectable marker in gene transfer to progenitor cells." Hum Gene Ther 14(4): 403-12.

Vahdat, L. T., K. Papadopoulos, C. Balmaceda, T. McGovern, J. Dunleavy, E. Kaufman, et al. (1998). "Phase I trial of sequential high-dose chemotherapy with escalating dose paclitaxel, melphalan, and cyclophosphamide, thiotepa, and carboplatin with peripheral blood progenitor support in women with responding metastatic breast cancer." Clin Cancer Res 4(7): 1689-95.

Valsesia-Wittmann, S., F. J. Morling, B. H. Nilson, Y. Takeuchi, S. J. Russell and F. L. Cosset (1996). "Improvement of retroviral retargeting by using amino acid spacers between an additional binding domain and the N terminus of Moloney murine leukemia virus SU." J Virol 70(3): 2059-64.

van Zeijl, M., S. V. Johann, E. Closs, J. Cunningham, R. Eddy, T. B. Shows, et al. (1994). "A human amphotropic retrovirus receptor is a second member of the gibbon ape leukemia virus receptor family." Proc Natl Acad Sci U S A 91(3): 1168-72.

von Laer, D., A. Corovic, B. Vogt, U. Herwig, S. Roscher, B. Fehse, et al. (2000). "Amphotropic and VSV-G-pseudotyped retroviral vectors transduce human

178 hematopoietic progenitor cells with similar efficiency." Bone Marrow Transplant 25 Suppl 2: S75-9.

Vowells, S. J., S. Sekhsaria, H. L. Malech, M. Shalit and T. A. Fleisher (1995). "Flow cytometric analysis of the granulocyte respiratory burst: a comparison study of fluorescent probes." J Immunol Methods 178(1): 89-97.

Wang, H., M. P. Kavanaugh, R. A. North and D. Kabat (1991). "Cell-surface receptor for ecotropic murine retroviruses is a basic amino-acid transporter." Nature 352(6337): 729- 31.

Ward, M., R. Sattler, I. R. Grossman, A. J. Bell, Jr., D. Skerrett, L. Baxi, et al. (2003). "A stable murine-based RD114 retroviral packaging line efficiently transduces human hematopoietic cells." Mol Ther 8(5): 804-12.

Warnakulasuriyarachchi, D., S. Cerquozzi, H. H. Cheung and M. Holcik (2004). "Translational induction of the inhibitor of apoptosis protein HIAP2 during endoplasmic reticulum stress attenuates cell death and is mediated via an inducible internal ribosome entry site element." J Biol Chem 279(17): 17148-57 Epub 2004 Feb 11.

Woods, N. B., A. Muessig, M. Schmidt, J. Flygare, K. Olsson, P. Salmon, et al. (2003). "Lentiviral vector transduction of NOD/SCID repopulating cells results in multiple vector integrations per transduced cell: risk of insertional mutagenesis." Blood 101(4): 1284-9.

Wu, B. W., J. Lu, T. K. Gallaher, W. F. Anderson and P. M. Cannon (2000). "Identification of regions in the Moloney murine leukemia virus SU protein that tolerate the insertion of an integrin-binding peptide." Virology 269(1): 7-17.

Xu-Welliver, M., S. Kanugula and A. E. Pegg (1998). "Isolation of human O6- alkylguanine-DNA alkyltransferase mutants highly resistant to inactivation by O6- benzylguanine." Cancer Res 58(9): 1936-45.

Yu, S. F., T. von Ruden, P. W. Kantoff, C. Garber, M. Seiberg, U. Ruther, et al. (1986). "Self-inactivating retroviral vectors designed for transfer of whole genes into mammalian cells." Proc Natl Acad Sci U S A 83(10): 3194-8.

179 Yu, X., X. Zhan, J. D'Costa, V. M. Tanavde, Z. Ye, T. Peng, et al. (2003). "Lentiviral vectors with two independent internal promoters transfer high-level expression of multiple transgenes to human hematopoietic stem-progenitor cells." Mol Ther 7(6): 827- 38.

Zavorotinskaya, T. and L. M. Albritton (1999). "A hydrophobic patch in ecotropic murine leukemia virus envelope protein is the putative binding site for a critical tyrosine residue on the cellular receptor." J Virol 73(12): 10164-72.

Zavorotinskaya, T. and L. M. Albritton (1999). "Suppression of a fusion defect by second site mutations in the ecotropic murine leukemia virus surface protein." J Virol 73(6): 5034-5042.

Zennou, V., C. Petit, D. Guetard, U. Nerhbass, L. Montagnier and P. Charneau (2000). "HIV-1 genome nuclear import is mediated by a central DNA flap." Cell 101(2): 173-85.

Zhang, B., H. Q. Xia, G. Cleghorn, G. Gobe, M. West and M. Q. Wei (2001). "A highly efficient and consistent method for harvesting large volumes of high-titre lentiviral vectors." Gene Ther 8(22): 1745-51.

Zhao, S., M. A. Weinreich, K. Ihara, R. E. Richard and C. A. Blau (2004). "In vivo selection of genetically modified erythroid cells using a jak2-based cell growth switch." Mol Ther 10(3): 456-68.

Zhao, S. C., M. X. Li, D. Banerjee, B. I. Schweitzer, S. Mineishi, E. Gilboa, et al. (1994). "Long-term protection of recipient mice from lethal doses of methotrexate by marrow infected with a double-copy vector retrovirus containing a mutant dihydrofolate reductase." Cancer Gene Ther 1(1): 27-33.

Zhen, L., A. A. King, Y. Xiao, S. J. Chanock, S. H. Orkin and M. C. Dinauer (1993). "Gene targeting of X chromosome-linked chronic granulomatous disease locus in a human myeloid leukemia cell line and rescue by expression of recombinant gp91phox." Proc Natl Acad Sci U S A 90(21): 9832-9836.

Zhou, S., J. J. Morris, Y. Barnes, L. Lan, J. D. Schuetz and B. P. Sorrentino (2002). "Bcrp1 gene expression is required for normal numbers of side population stem cells in mice, and confers relative protection to mitoxantrone in hematopoietic cells in vivo." Proc Natl Acad Sci U S A 99(19): 12339-44 Epub 2002 Sep 06.

180

Zhou, S., J. D. Schuetz, K. D. Bunting, A. M. Colapietro, J. Sampath, J. J. Morris, et al. (2001). "The ABC transporter Bcrp1/ABCG2 is expressed in a wide variety of stem cells and is a molecular determinant of the side-population phenotype." Nat Med 7(9): 1028- 34.

Zhou, Y., J. Aran, M. M. Gottesman and I. Pastan (1998). "Co-expression of human adenosine deaminase and multidrug resistance using a bicistronic retroviral vector." Hum Gene Ther 9(3): 287-93.

Zhu, N. L., P. M. Cannon, D. Chen and W. F. Anderson (1998). "Mutational analysis of the fusion peptide of Moloney murine leukemia virus transmembrane protein p15E." J Virol 72(2): 1632-9.

Zhu, Y., G. Feuer, S. L. Day, S. Wrzesinski and V. Planelles (2001). "Multigene lentiviral vectors based on differential splicing and translational control." Mol Ther 4(4): 375-82.

Zielske, S. P. and S. L. Gerson (2003). "Cytokines, including stem cell factor alone, enhance lentiviral transduction in nondividing human LTCIC and NOD/SCID repopulating cells." Mol Ther 7(3): 325-33.

Zielske, S. P., J. S. Reese, K. T. Lingas, J. R. Donze and S. L. Gerson (2003). "In vivo selection of MGMT(P140K) lentivirus-transduced human NOD/SCID repopulating cells without pretransplant irradiation conditioning." J Clin Invest 112(10): 1561-70.

Zufferey, R., T. Dull, R. J. Mandel, A. Bukovsky, D. Quiroz, L. Naldini, et al. (1998). "Self-inactivating lentivirus vector for safe and efficient in vivo gene delivery." J Virol 72(12): 9873-80.

181